This is the multi-page printable view of this section. Click here to print.

Return to the regular view of this page.

Machine learning functions

Machine learning functions let you work with your data set in different stages of the data analysis process:.

1: Data preparation

1.1: BALANCE
1.2: CHI_SQUARED
1.3: CORR_MATRIX
1.4: DETECT_OUTLIERS
1.5: IFOREST
1.6: IMPUTE
1.7: NORMALIZE
1.8: NORMALIZE_FIT
1.9: ONE_HOT_ENCODER_FIT
1.10: PCA
1.11: SUMMARIZE_CATCOL
1.12: SUMMARIZE_NUMCOL
1.13: SVD

2: Machine learning algorithms

2.1: ARIMA
2.2: AUTOREGRESSOR
2.3: BISECTING_KMEANS
2.4: KMEANS
2.5: KPROTOTYPES
2.6: LINEAR_REG
2.7: LOGISTIC_REG
2.8: MOVING_AVERAGE
2.9: NAIVE_BAYES
2.10: POISSON_REG
2.11: RF_CLASSIFIER
2.12: RF_REGRESSOR
2.13: SVM_CLASSIFIER
2.14: SVM_REGRESSOR
2.15: XGB_CLASSIFIER
2.16: XGB_REGRESSOR

3: Model evaluation

3.1: CONFUSION_MATRIX
3.2: CROSS_VALIDATE
3.3: ERROR_RATE
3.4: LIFT_TABLE
3.5: MSE
3.6: PRC
3.7: READ_TREE
3.8: RF_PREDICTOR_IMPORTANCE
3.9: ROC
3.10: RSQUARED
3.11: XGB_PREDICTOR_IMPORTANCE

4: Model management

4.1: CHANGE_MODEL_STATUS
4.2: EXPORT_MODELS
4.3: GET_MODEL_ATTRIBUTE
4.4: GET_MODEL_SUMMARY
4.5: IMPORT_MODELS
4.6: REGISTER_MODEL
4.7: UPGRADE_MODEL

5: Transformation functions

5.1: APPLY_BISECTING_KMEANS
5.2: APPLY_IFOREST
5.3: APPLY_INVERSE_PCA
5.4: APPLY_INVERSE_SVD
5.5: APPLY_KMEANS
5.6: APPLY_KPROTOTYPES
5.7: APPLY_NORMALIZE
5.8: APPLY_ONE_HOT_ENCODER
5.9: APPLY_PCA
5.10: APPLY_SVD
5.11: PREDICT_ARIMA
5.12: PREDICT_AUTOREGRESSOR
5.13: PREDICT_LINEAR_REG
5.14: PREDICT_LOGISTIC_REG
5.15: PREDICT_MOVING_AVERAGE
5.16: PREDICT_NAIVE_BAYES
5.17: PREDICT_NAIVE_BAYES_CLASSES
5.18: PREDICT_PMML
5.19: PREDICT_POISSON_REG
5.20: PREDICT_RF_CLASSIFIER
5.21: PREDICT_RF_CLASSIFIER_CLASSES
5.22: PREDICT_RF_REGRESSOR
5.23: PREDICT_SVM_CLASSIFIER
5.24: PREDICT_SVM_REGRESSOR
5.25: PREDICT_TENSORFLOW
5.26: PREDICT_TENSORFLOW_SCALAR
5.27: PREDICT_XGB_CLASSIFIER
5.28: PREDICT_XGB_CLASSIFIER_CLASSES
5.29: PREDICT_XGB_REGRESSOR
5.30: REVERSE_NORMALIZE

Machine learning functions let you work with your data set in different stages of the data analysis process:

Preparing models
Training models
Evaluating models
Applying models
Managing models

Some Vertica machine learning functions are implemented as Vertica UDx functions, while others are implemented as meta-functions:

A UDx function accepts an input relation name from a FROM clause. The SELECT statement that calls the functions is composable—it can be used as a sub-query in another SELECT statement.
A meta-function accepts the input relation name as a single-quoted string passed to it as an argument or a named parameter. The data that the SELECT statement returns cannot be used in a sub-query. Machine learning meta-functions do not support temporary tables.

All machine learning functions automatically cast NUMERIC arguments to FLOAT.

Important

Before using a machine learning function, be aware that any open transaction on the current session might be committed.

1 - Data preparation

Vertica supports machine learning functions that prepare data as needed before subjecting it to analysis.

1.1 - BALANCE

Returns a view with an equal distribution of the input data based on the response_column.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

BALANCE ( 'output-view', 'input-relation', 'response-column', 'balance-method'
       [ USING PARAMETERS sampling_ratio=ratio ] )

Arguments

output-view

The name of the view where Vertica saves the balanced data from the input relation.

Note

Note: The view that results from this function employs a random function. Its content can differ each time it is used in a query. To make the operations on the view predictable, store it in a regular table.

input-relation

The table or view that contains the data the function uses to create a more balanced data set. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.

response-column

Name of the input column that represents the dependent variable, of type VARCHAR or INTEGER.

balance-method

Specifies a method to select data from the minority and majority classes, one of the following.

hybrid_sampling: Performs over-sampling and under-sampling on different classes so each class is equally represented.
over_sampling: Over-samples on all classes, with the exception of the most majority class, towards the most majority class's cardinality.
under_sampling: Under-samples on all classes, with the exception of the most minority class, towards the most minority class's cardinality.
weighted_sampling: An alias of under_sampling.

Parameters

ratio: The desired ratio between the majority class and the minority class. This value has no effect when used with balance method hybrid_sampling.
Default: 1.0

Privileges

Non-superusers:

SELECT privileges on the input relation
CREATE privileges on the output view schema

Examples

=> CREATE TABLE backyard_bugs (id identity, bug_type int, finder varchar(20));
CREATE TABLE

=> COPY backyard_bugs FROM STDIN;
Enter data to be copied followed by a newline.
End with a backslash and a period on a line by itself.
>> 1|Ants
>> 1|Beetles
>> 3|Ladybugs
>> 3|Ants
>> 3|Beetles
>> 3|Caterpillars
>> 2|Ladybugs
>> 3|Ants
>> 3|Beetles
>> 1|Ladybugs
>> 3|Ladybugs
>> \.

=> SELECT bug_type, COUNT(bug_type) FROM backyard_bugs GROUP BY bug_type;
 bug_type | COUNT
----------+-------
        2 |     1
        1 |     3
        3 |     7
(3 rows)

=> SELECT BALANCE('backyard_bugs_balanced', 'backyard_bugs', 'bug_type', 'under_sampling');
         BALANCE
--------------------------
 Finished in 1 iteration

(1 row)

=> SELECT bug_type, COUNT(bug_type) FROM backyard_bugs_balanced GROUP BY bug_type;
----------+-------
        2 |     1
        1 |     2
        3 |     1
(3 rows)

1.2 - CHI_SQUARED

Computes the conditional chi-Square independence test on two categorical variables to find the likelihood that the two variables are independent.

Computes the conditional chi-square independence test on two categorical variables to find the likelihood that the two variables are independent. To condition the independence test on another set of variables, you can partition the data on these variables using a PARTITION BY clause.

Tip

If a categorical column is not of a numeric data type, you can use the HASH function to convert it into a column of type INT, where each category is mapped to a unique integer. However, note that NULL values are hashed to zero, so they will be included in the test instead of skipped by the function.

This function is a multi-phase transform function.

Syntax

CHI_SQUARED( 'x-column', 'y-column' 
    [ USING PARAMETERS param=value[,...] ] )

Arguments

x-column, y-column: Columns in the input relation to be tested for dependency with each other. These columns must contain categorical data in numeric format.

Parameters

x_cardinality: Integer in the range [1, 20], the cardinality of x-column. If the cardinality of x-column is less than the default value of 20, setting this parameter can decrease the amount memory used by the function.
Default: 20
y_cardinality: Integer in the range [1, 20], the cardinality of y-column. If the cardinality of y-column is less than the default value of 20, setting this parameter can decrease the amount memory used by the function.
Default: 20
alpha: Float in the range (0.0, 1.0), the significance level. If the returned pvalue is less than this value, the null hypothesis, which assumes the variables are independent, is rejected.
Default: 0.05

Returns

The function returns two values:

pvalue (float): the confidence that the two variables are independent. If this value is greater than the alpha parameter value, the null hypothesis is accepted and the variables are considered independent.
independent (boolean): true if the variables are independent; otherwise, false.

Privileges

SELECT privileges on the input relation

Examples

The following examples use the titanic dataset from the machine learning example data. If you have not downloaded these datasets, see Download the machine learning example data for instructions.

The titanic_training table contains data related to passengers on the Titanic, including:

pclass: the ticket class of the passenger, ranging from 1st class to 3rd class
survived: whether the passenger survived, where 1 is yes and 0 is no
gender: gender of the passenger
sibling_and_spouse_count: number of siblings aboard the Titanic
embarkation_point: port of embarkation

To test whether the survival of a passenger is dependent on their ticket class, run the following chi-square test:

=> SELECT CHI_SQUARED(pclass, survived USING PARAMETERS x_cardinality=3, y_cardinality=2, alpha=0.05) OVER() FROM titanic_training;
 pvalue | independent
--------+-------------
      0 | f
(1 row)

With a returned pvalue of zero, the null hypothesis is rejected and you can conclude that the survived and pclass variables are dependent. To test whether this outcome is conditional on the gender of the passenger, partition by the gender column in the OVER clause:

=> SELECT CHI_SQUARED(pclass, survived USING PARAMETERS x_cardinality=3, y_cardinality=2) OVER(PARTITION BY gender) FROM titanic;
 pvalue | independent
--------+-------------
      0 | f
(1 row)

As the pvalue is still zero, it is clear that the dependence of the pclass and survived variables is not conditional on the gender of the passenger.

If one of the categorical columns that you want to test is not a numeric type, use the HASH function to convert it into type INT:

=> SELECT CHI_SQUARED(sibling_and_spouse_count, HASH(embarkation_point) USING PARAMETERS alpha=0.05) OVER() FROM titanic_training;
       pvalue       | independent
--------------------+-------------
 0.0753039994044853 | t
(1 row)

The returned pvalue is greater than alpha, meaning the null hypothesis is accepted and the sibling_and_spouse_count and embarkation_point are independent.

1.3 - CORR_MATRIX

Takes an input relation with numeric columns, and calculates the Pearson Correlation Coefficient between each pair of its input columns.

Takes an input relation with numeric columns, and calculates the Pearson Correlation Coefficient between each pair of its input columns. The function is implemented as a Multi-Phase Transform function.

Syntax

CORR_MATRIX ( input-columns ) OVER()

Arguments

input-columns: A comma-separated list of the columns in the input table. The input columns can be of any numeric type or BOOL, but they will be converted internally to FLOAT. The number of input columns must be more than 1 and not more than 1600.

Returns

CORR_MATRIX returns the correlation matrix in triplet format. That is, each pair-wise correlation is identified by three returned columns: name of the first variable, name of the second variable, and the correlation value of the pair. The function also returns two extra columns: number_of_ignored_input_rows and number_of_processed_input_rows. The value of the fourth/fifth column indicates the number of rows from the input which are ignored/used to calculate the corresponding correlation value. Any input pair with NULL, Inf, or NaN is ignored.

The correlation matrix is symmetric with a value of 1 on all diagonal elements; therefore, it can return only the value of elements above the diagonals—that is, the upper triangle. Nevertheless, the function returns the entire matrix to simplify any later operations. Then, the number of output rows is:

(#input-columns)^2

The first two output columns are of type VARCHAR(128), the third one is of type FLOAT, and the last two are of type INT.

Notes

The contents of the OVER clause must be empty.
The function returns no rows when the input table is empty.
When any of X_i and Y_i is NULL, Inf, or NaN, the pair will not be included in the calculation of CORR(X, Y). That is, any input pair with NULL, Inf, or NaN is ignored.
For the pair of (X,X), regardless of the contents of X: CORR(X,X) = 1, number_of_ignored_input_rows = 0, and number_of_processed_input_rows = #input_rows.
When (NSUMX2 == SUMXSUMX) or (NSUMY2 == SUMYSUMY) then value of CORR(X, Y) will be NULL. In theory it can happen in case of a column with constant values; nevertheless, it may not be always observed because of rounding error.
In the special case where all pair values of (X_i,Y_i) contain NULL, inf, or NaN, and X != Y: CORR(X,Y)=NULL.

Examples

The following example

SELECT variable_name_1 | variable_name_2 ----------------+------------- Sepal.Length    | Sepal.Width Sepal.Width     | Sepal.Length Sepal.Length    | Petal.Length Petal.Length    | Sepal.Length Sepal.Length    | Petal.Width Petal.Width     | Sepal.Length Sepal.Width     | Petal.Length Petal.Length    | Sepal.Width Sepal.Width     | Petal.Width Petal.Width     | Sepal.Width Petal.Length    | Petal.Width Petal.Width     | Petal.Length Sepal.Length    | Sepal.Length    |1 Sepal.Width     | Sepal.Width     |1 Petal.Length    | Petal.Length    |1 Petal.Width     | Petal.Width     |1 (16 rows)

uses the iris dataset.* CORR_MATRIX("Sepal.Length", "Sepal.Width", "Petal.Length", "Petal.Width") OVER() FROM iris; | corr_value | number_of_ignored_input_rows | number_of_processed_input_rows ----+-------------------+------------------------------+-------------------------------- |-0.117569784133002 | 0 | 150 |-0.117569784133002 | 0 | 150 |0.871753775886583 | 0 | 150 |0.871753775886583 | 0 | 150 |0.817941126271577 | 0 | 150 |0.817941126271577 | 0 | 150 |-0.42844010433054 | 0 | 150 |-0.42844010433054 | 0 | 150 |-0.366125932536439 | 0 | 150 |-0.366125932536439 | 0 | 150 |0.962865431402796 | 0 | 150 |0.962865431402796 | 0 | 150 | 0 | 150 | 0 | 150 | 0 | 150 | 0 | 150

1.4 - DETECT_OUTLIERS

Returns the outliers in a data set based on the outlier threshold.

Returns the outliers in a data set based on the outlier threshold. The output is a table that contains the outliers. DETECT_OUTLIERS uses the detection method robust_szcore to normalize each input column. The function then identifies as outliers all rows that contain a normalized value greater than the default or specified threshold.

Note

You can calculate normalized column values with Vertica functions NORMALIZE and NORMALIZE_FIT.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

DETECT_OUTLIERS ( 'output-table', 'input-relation','input-columns', 'detection-method'
        [ USING PARAMETERS
              [outlier_threshold = threshold]
              [, exclude_columns = 'excluded-columns']
              [, partition_columns = 'partition-columns'] ] )

Arguments

output-table: The name of the table where Vertica saves rows that are outliers along the chosen input_columns. All columns are present in this table.
input-relation: The table or view that contains outlier data. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.
input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. Input columns must be of type numeric.
detection-method: The outlier detection method to use, set to robust_zscore.

Parameters

outlier_threshold: The minimum normalized value in a row that is used to identify that row as an outlier.
Default: 3.0
exclude_columns: Comma-separated list of column names from input-columns to exclude from processing.
partition_columns: Comma-separated list of column names from the input table or view that defines the partitions. DETECT_OUTLIERS detects outliers among each partition separately.
Default: empty list

Privileges

Non-superusers:

SELECT privileges on the input relation
CREATE privileges on the output table

Examples

The following example shows how to use DETECT_OUTLIERS:

=> CREATE TABLE baseball_roster (id identity, last_name varchar(30), hr int, avg float);
CREATE TABLE

=> COPY baseball_roster FROM STDIN;
Enter data to be copied followed by a newline.
End with a backslash and a period on a line by itself.
>> Polo|7|.233
>> Gloss|45|.170
>> Gus|12|.345
>> Gee|1|.125
>> Laus|3|.095
>> Hilltop|16|.222
>> Wicker|78|.333
>> Scooter|0|.121
>> Hank|999999|.8888
>> Popup|35|.378
>> \.


=> SELECT * FROM baseball_roster;
 id | last_name |   hr   |  avg
----+-----------+--------+--------
  3 | Gus       |     12 |  0.345
  4 | Gee       |      1 |  0.125
  6 | Hilltop   |     16 |  0.222
 10 | Popup     |     35 |  0.378
  1 | Polo      |      7 |  0.233
  7 | Wicker    |     78 |  0.333
  9 | Hank      | 999999 | 0.8888
  2 | Gloss     |     45 |   0.17
  5 | Laus      |      3 |  0.095
  8 | Scooter   |      0 |  0.121
(10 rows)

=> SELECT DETECT_OUTLIERS('baseball_outliers', 'baseball_roster', 'id, hr, avg', 'robust_zscore' USING PARAMETERS
outlier_threshold=3.0);

     DETECT_OUTLIERS
--------------------------
 Detected 2 outliers

(1 row)

=> SELECT * FROM baseball_outliers;
 id | last_name | hr         | avg
----+-----------+------------+-------------
  7 | Wicker    |         78 |       0.333
  9 | Hank      |     999999 |      0.8888
(2 rows)

1.5 - IFOREST

Trains and returns an isolation forest (iForest) model.

Trains and returns an isolation forest (iForest) model. After you train the model, you can use the APPLY_IFOREST function to predict outliers in an input relation.

For more information about how the iForest algorithm works, see Isolation Forest.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

IFOREST( 'model-name', 'input-relation', 'input-columns' [ USING PARAMETERS param=value[,...] ] )

Arguments

model-name: Identifies the model to create, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.
input-relation: The table or view that contains the input data for IFOREST.
input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. Columns must be of types CHAR, VARCHAR, BOOL, INT, or FLOAT.
Columns of types CHAR, VARCHAR, and BOOL are treated as categorical features; all others are treated as numeric features.

Parameters

exclude_columns: Comma-separated list of column names from input-columns to exclude from processing.
Default: Empty string ('')
ntree: Integer in the range [1, 1000], specifies the number of trees in the forest.
Default: 100
sampling_size: Float in the range (0.0, 1.0], specifies the portion of the input data set that is randomly picked, without replacement, for training each tree.
Default: 0.632
col_sample_by_tree: Float in the range (0.0, 1.0], specifies the fraction of columns that are randomly picked for training each tree.
Default: 1.0
max_depth: Integer in the range [1, 100], specifies the maximum depth for growing each tree.
Default: 10
nbins: Integer in the range [2, 1000], specifies the number of bins used to discretize continuous features.
Default: 32

Model Attributes

details

Details about the function's predictor columns, including:

predictor: Names of the predictors in the same order specified when training the model.
type: Types of the predictors in the same order as their names in predictor.

tree_count

Number of trees in the model.

rejected_row_count

Number of rows in input-relation that were skipped because they contained an invalid value.

accepted_row_count

Total number of rows in input-relation minus rejected_row_count.

call_string

Value of all input arguments that were specified at the time the function was called.

Privileges

Non-superusers:

CREATE privileges on the schema where the model is created
SELECT privileges on the input relation

Examples

In the following example, the input data to the function contains columns of type INT, VARCHAR, and FLOAT:

=> SELECT IFOREST('baseball_anomalies','baseball','team, hr, hits, avg, salary' USING PARAMETERS ntree=75, sampling_size=0.7,
max_depth=15);
IFOREST
----------
Finished
(1 row)

You can verify that all the input columns were read in correctly by calling GET_MODEL_SUMMARY and checking the details section:

=> SELECT GET_MODEL_SUMMARY(USING PARAMETERS model_name='baseball_anomalies');
GET_MODEL_SUMMARY
-------------------------------------------------------------------------------------------------------------------------------------

===========
call_string
===========
SELECT iforest('public.baseball_anomalies', 'baseball', 'team, hr, hits, avg, salary' USING PARAMETERS exclude_columns='', ntree=75,
sampling_size=0.7, col_sample_by_tree=1, max_depth=15, nbins=32);

=======
details
=======
predictor|      type
---------+----------------
  team   |char or varchar
   hr    |      int
  hits   |      int
   avg   |float or numeric
 salary  |float or numeric


===============
Additional Info
===============
       Name       |Value
------------------+-----
    tree_count    | 75
rejected_row_count|  0
accepted_row_count|1000

(1 row)

1.6 - IMPUTE

Imputes missing values in a data set with either the mean or the mode, based on observed values for a variable in each column.

Imputes missing values in a data set with either the mean or the mode, based on observed values for a variable in each column. This function supports numeric and categorical data types.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

IMPUTE( 'output-view', 'input-relation', 'input-columns', 'method'
        [ USING PARAMETERS [exclude_columns = 'excluded-columns'] [, partition_columns = 'partition-columns'] ] )

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Arguments

output-view

Name of the view that shows the input table with imputed values in place of missing values. In this view, rows without missing values are kept intact while the rows with missing values are modified according to the specified method.

input-relation

The table or view that contains the data for missing-value imputation. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.

input-columns

Comma-separated list of input columns where missing values will be replaced, or asterisk (*) to specify all columns. All columns must be of type numeric or BOOLEAN.

method

The method to compute the missing value replacements, one of the following:

mean: The missing values in each column will be replaced by the mean of that column. This method can be used for numeric data only.
mode: The missing values in each column will be replaced by the most frequent value in that column. This method can be used for categorical data only.

Parameters

exclude_columns: Comma-separated list of column names from input-columns to exclude from processing.
partition_columns: Comma-separated list of column names from the input relation that defines the partitions.

Privileges

Non-superusers:

SELECT privileges on the input relation
CREATE privileges on the output view schema

Examples

Execute IMPUTE on the small_input_impute table, specifying the mean method:

=> SELECT impute('output_view','small_input_impute', 'pid, x1,x2,x3,x4','mean'
USING PARAMETERS exclude_columns='pid');
impute
--------------------------
Finished in 1 iteration
(1 row)

Execute IMPUTE, specifying the mode method:

=> SELECT impute('output_view3','small_input_impute', 'pid, x5,x6','mode' USING PARAMETERS exclude_columns='pid');
impute
--------------------------
Finished in 1 iteration
(1 row)

1.7 - NORMALIZE

Runs a normalization algorithm on an input relation.

Runs a normalization algorithm on an input relation. The output is a view with the normalized data.

Note

Note: This function differs from NORMALIZE_FIT, which creates and stores a model rather than creating a view definition. This can lead to different performance characteristics between the two functions.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

NORMALIZE ( 'output-view', 'input-relation', 'input-columns', 'normalization-method'
           [ USING PARAMETERS exclude_columns = 'excluded-columns' ] )

Arguments

output-view

The name of the view showing the input relation with normalized data replacing the specified input columns. .

input-relation

The table or view that contains the data to normalize. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.

input-columns

Comma-separated list of numeric input columns that contain the values to normalize, or asterisk (*) to select all columns.

normalization-method

The normalization method to use, one of the following:

minmax
zscore
robust_zscore

If infinity values appear in the table, the method ignores those values.

Parameters

exclude_columns: Comma-separated list of column names from input-columns to exclude from processing.

Privileges

Non-superusers:

SELECT privileges on the input relation
CREATE privileges on the output view schema

Examples

These examples show how you can use the NORMALIZE function on the wt and hp columns in the mtcars table.

Execute the NORMALIZE function, and specify the minmax method:

=> SELECT NORMALIZE('mtcars_norm', 'mtcars',
                    'wt, hp', 'minmax');
        NORMALIZE
--------------------------
 Finished in 1 iteration

(1 row)

Execute the NORMALIZE function, and specify the zscore method:

=> SELECT NORMALIZE('mtcars_normz','mtcars',
                    'wt, hp', 'zscore');
        NORMALIZE
--------------------------
 Finished in 1 iteration

(1 row)

Execute the NORMALIZE function, and specify the robust_zscore method:

=> SELECT NORMALIZE('mtcars_normz', 'mtcars',
                    'wt, hp', 'robust_zscore');
        NORMALIZE
--------------------------
 Finished in 1 iteration

(1 row)

1.8 - NORMALIZE_FIT

This function differs from NORMALIZE, which directly outputs a view with normalized results, rather than storing normalization parameters into a model for later operation.

Note

This function differs from NORMALIZE, which directly outputs a view with normalized results, rather than storing normalization parameters into a model for later operation.

NORMALIZE_FIT computes normalization parameters for each of the specified columns in an input relation. The resulting model stores the normalization parameters. For example, for MinMax normalization, the minimum and maximum value of each column are stored in the model. The generated model serves as input to functions APPLY_NORMALIZE and REVERSE_NORMALIZE.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

NORMALIZE_FIT ( 'model-name', 'input-relation', 'input-columns', 'normalization-method'
        [ USING PARAMETERS  [exclude_columns = 'excluded-columns'] [, output_view = 'output-view'] ] )

Arguments

model-name

Identifies the model to create, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.

input-relation

input-columns

Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. Input columns must be of data type numeric.

normalization-method

The normalization method to use, one of the following:

minmax
zscore
robust_zscore

If you specify robust_zscore, NORMALIZE_FIT uses the function APPROXIMATE_MEDIAN [aggregate].

All normalization methods ignore infinity, negative infinity, or NULL values in the input relation.

Parameters

exclude_columns: Comma-separated list of column names from input-columns to exclude from processing.
output_view: Name of the view that contains all columns from the input relation, with the specified input columns normalized.

Model attributes

data

Normalization method set to minmax:

colNames: Model column names
mins: Minimum value of each column
maxes: Maximum value of each column

Privileges

Non-superusers:

CREATE privileges on the schema where the model is created
SELECT privileges on the input relation
CREATE privileges on the output view schema

Examples

The following example creates a model with NORMALIZE_FIT using the wt and hp columns in table mtcars , and then uses this model in successive calls to APPLY_NORMALIZE and REVERSE_NORMALIZE.

=> SELECT NORMALIZE_FIT('mtcars_normfit', 'mtcars', 'wt,hp', 'minmax');
NORMALIZE_FIT
---------------
Success
(1 row)

The following call to APPLY_NORMALIZE specifies the hp and cyl columns in table mtcars, where hp is in the normalization model and cyl is not in the normalization model:

=> CREATE TABLE mtcars_normalized AS SELECT APPLY_NORMALIZE (hp, cyl USING PARAMETERS model_name = 'mtcars_normfit') FROM mtcars;
CREATE TABLE
=> SELECT * FROM mtcars_normalized;
          hp        | cyl
--------------------+-----
  0.434628975265018 | 8
  0.681978798586572 | 8
  0.434628975265018 | 6
                  1 | 8
  0.540636042402827 | 8
                  0 | 4
  0.681978798586572 | 8
 0.0459363957597173 | 4
  0.434628975265018 | 8
  0.204946996466431 | 6
  0.250883392226148 | 6
  0.049469964664311 | 4
  0.204946996466431 | 6
  0.201413427561837 | 4
  0.204946996466431 | 6
  0.250883392226148 | 6
  0.049469964664311 | 4
  0.215547703180212 | 4
 0.0353356890459364 | 4
  0.187279151943463 | 6
  0.452296819787986 | 8
  0.628975265017668 | 8
  0.346289752650177 | 8
  0.137809187279152 | 4
  0.749116607773852 | 8
  0.144876325088339 | 4
  0.151943462897526 | 4
  0.452296819787986 | 8
  0.452296819787986 | 8
  0.575971731448763 | 8
  0.159010600706714 | 4
  0.346289752650177 | 8
(32 rows)

=> SELECT REVERSE_NORMALIZE (hp, cyl USING PARAMETERS model_name='mtcars_normfit') FROM mtcars_normalized;
  hp | cyl
-----+-----
 175 | 8
 245 | 8
 175 | 6
 335 | 8
 205 | 8
  52 | 4
 245 | 8
  65 | 4
 175 | 8
 110 | 6
 123 | 6
  66 | 4
 110 | 6
 109 | 4
 110 | 6
 123 | 6
  66 | 4
 113 | 4
  62 | 4
 105 | 6
 180 | 8
 230 | 8
 150 | 8
  91 | 4
 264 | 8
  93 | 4
  95 | 4
 180 | 8
 180 | 8
 215 | 8
  97 | 4
 150 | 8
(32 rows)

The following call to REVERSE_NORMALIZE also specifies the hp and cyl columns in table mtcars, where hp is in normalization model mtcars_normfit, and cyl is not in the normalization model.

=> SELECT REVERSE_NORMALIZE (hp, cyl USING PARAMETERS model_name='mtcars_normfit') FROM mtcars_normalized;
       hp        | cyl
-----------------+-----
205.000005722046 |   8
150.000000357628 |   8
150.000000357628 |   8
93.0000016987324 |   4
 174.99999666214 |   8
94.9999992102385 |   4
214.999997496605 |   8
97.0000009387732 |   4
245.000006556511 |   8
 174.99999666214 |   6
             335 |   8
245.000006556511 |   8
62.0000002086163 |   4
 174.99999666214 |   8
230.000002026558 |   8
              52 |   4
263.999997675419 |   8
109.999999523163 |   6
123.000002324581 |   6
64.9999996386468 |   4
66.0000005029142 |   4
112.999997898936 |   4
109.999999523163 |   6
180.000000983477 |   8
180.000000983477 |   8
108.999998658895 |   4
109.999999523163 |   6
104.999999418855 |   6
123.000002324581 |   6
180.000000983477 |   8
66.0000005029142 |   4
90.9999999701977 |   4
(32 rows)

1.9 - ONE_HOT_ENCODER_FIT

Generates a sorted list of each of the category levels for each feature to be encoded, and stores the model.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

ONE_HOT_ENCODER_FIT ( 'model-name', 'input-relation','input-columns'
        [ USING PARAMETERS
              [exclude_columns = 'excluded-columns']
              [, output_view = 'output-view']
              [, extra_levels = 'category-levels'] ] )

Arguments

model-name: Identifies the model to create, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.
input-relation: The table or view that contains the data for one hot encoding. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.
input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. Input columns must be INTEGER, BOOLEAN, VARCHAR, or dates.

Parameters

exclude_columns: Comma-separated list of column names from input-columns to exclude from processing.
output_view: The name of the view that stores the input relation and the one hot encodings. Columns are returned in the order they appear in the input relation, with the one-hot encoded columns appended after the original columns.
extra_levels: Additional levels in each category that are not in the input relation. This parameter should be passed as a string that conforms with the JSON standard, with category names as keys, and lists of extra levels in each category as values.

Model attributes

call_string

The value of all input arguments that were specified at the time the function was called.

varchar_categories integer_categories boolean_categories date_categories

Settings for all:

category_name: Column name
category_level: Levels of the category, sorted for each category
category_level_index: Index of this categorical level in the sorted list of levels for the category.

Privileges

Non-superusers:

CREATE privileges on the schema where the model is created
SELECT privileges on the input relation
CREATE privileges on the output view schema

Examples

=> SELECT ONE_HOT_ENCODER_FIT ('one_hot_encoder_model','mtcars','*'
USING PARAMETERS exclude_columns='mpg,disp,drat,wt,qsec,vs,am');
ONE_HOT_ENCODER_FIT
--------------------
Success
(1 row)

1.10 - PCA

Computes principal components from the input table/view.

Computes principal components from the input table/view. The results are saved in a PCA model. Internally, PCA finds the components by using SVD on the co-variance matrix built from the input date. The singular values of this decomposition are also saved as part of the PCA model. The signs of all elements of a principal component could be flipped all together on different runs.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

PCA ( 'model-name', 'input-relation', 'input-columns'
        [ USING PARAMETERS
              [exclude_columns = 'excluded-columns']
              [, num_components = num-components]
              [, scale = is-scaled]
              [, method = 'method'] ] )

Arguments

model-name: Identifies the model to create, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.
input-relation: The table or view that contains the input data for PCA.
input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. All input columns must be a numeric data type.

Parameters

exclude_columns

Comma-separated list of column names from input-columns to exclude from processing.

num_components

The number of components to keep in the model. If this value is not provided, all components are kept. The maximum number of components is the number of non-zero singular values returned by the internal call to SVD. This number is less than or equal to SVD (number of columns, number of rows).

scale

A Boolean value that specifies whether to standardize the columns during the preparation step:

True: Use a correlation matrix instead of a covariance matrix.
False (default)

method

The method used to calculate PCA, can be set to LAPACK.

Model attributes

columns

The information about columns from the input relation used for creating the PCA model:

index
name

singular_values

The information about singular values found. They are sorted in descending order:

index
value
explained_variance : percentage of the variance in data that can be attributed to this singular value
accumulated_explained_variance : percentage of the variance in data that can be retained if we drop all singular values after this current one

principal_components

The principal components corresponding to the singular values mentioned above:

index: indies of the elements in each component
PC1
PC2
...

counters

The information collected during training the model, stored as name-value pairs:

counter_name
- accepted_row_count: number of valid rows in the data
- rejected_row_count: number of invalid rows (having NULL, INF or NaN) in the data
- iteration_count: number of iterations, always 1 for the current implementation of PCA
counter_value

call_string

The function call that created the model.

Privileges

Non-superusers:

CREATE privileges on the schema where the model is created
SELECT privileges on the input relation

Examples


=> SELECT PCA ('pcamodel', 'world','country,HDI,em1970,em1971,em1972,em1973,em1974,em1975,em1976,em1977,
em1978,em1979,em1980,em1981,em1982,em1983,em1984 ,em1985,em1986,em1987,em1988,em1989,em1990,em1991,em1992,
em1993,em1994,em1995,em1996,em1997,em1998,em1999,em2000,em2001,em2002,em2003,em2004,em2005,em2006,em2007,
em2008,em2009,em2010,gdp1970,gdp1971,gdp1972,gdp1973,gdp1974,gdp1975,gdp1976,gdp1977,gdp1978,gdp1979,gdp1980,
gdp1981,gdp1982,gdp1983,gdp1984,gdp1985,gdp1986,gdp1987,gdp1988,gdp1989,gdp1990,gdp1991,gdp1992,gdp1993,
gdp1994,gdp1995,gdp1996,gdp1997,gdp1998,gdp1999,gdp2000,gdp2001,gdp2002,gdp2003,gdp2004,gdp2005,gdp2006,
gdp2007,gdp2008,gdp2009,gdp2010' USING PARAMETERS exclude_columns='HDI,country');
PCA
---------------------------------------------------------------
Finished in 1 iterations.
Accepted Rows: 96  Rejected Rows: 0
(1 row)
=> CREATE TABLE worldPCA AS SELECT
APPLY_PCA (HDI,country,em1970,em1971,em1972,em1973,em1974,em1975,em1976,em1977,em1978,em1979,
em1980,em1981,em1982,em1983,em1984 ,em1985,em1986,em1987,em1988,em1989,em1990,em1991,em1992,em1993,em1994,
em1995,em1996,em1997,em1998,em1999,em2000,em2001,em2002,em2003,em2004,em2005,em2006,em2007,em2008,em2009,
em2010,gdp1970,gdp1971,gdp1972,gdp1973,gdp1974,gdp1975,gdp1976,gdp1977,gdp1978,gdp1979,gdp1980,gdp1981,gdp1982,
gdp1983,gdp1984,gdp1985,gdp1986,gdp1987,gdp1988,gdp1989,gdp1990,gdp1991,gdp1992,gdp1993,gdp1994,gdp1995,
gdp1996,gdp1997,gdp1998,gdp1999,gdp2000,gdp2001,gdp2002,gdp2003,gdp2004,gdp2005,gdp2006,gdp2007,gdp2008,
gdp2009,gdp2010 USING PARAMETERS model_name='pcamodel', exclude_columns='HDI, country', key_columns='HDI,
country',cutoff=.3)OVER () FROM world;
CREATE TABLE

=> SELECT * FROM worldPCA;
HDI   |       country       |       col1
------+---------------------+-------------------
0.886 | Belgium             |  79002.2946705704
0.699 | Belize              | -25631.6670012556
0.427 | Benin               | -40373.4104598122
0.805 | Chile               | -16805.7940082156
0.687 | China               | -37279.2893141103
0.744 | Costa Rica          | -19505.5631231635
0.4   | Cote d'Ivoire       | -38058.2060339272
0.776 | Cuba                | -23724.5779612041
0.895 | Denmark             |  117325.594028813
0.644 | Egypt               | -34609.9941604549
...
(96 rows)

=> SELECT APPLY_INVERSE_PCA (HDI, country, col1
    USING PARAMETERS model_name = 'pcamodel', exclude_columns='HDI,country',
    key_columns = 'HDI, country') OVER () FROM worldPCA;
HDI  |       country       |      em1970       |      em1971       |      em1972      |      em1973      |
      em1974      |      em1975       |      em1976|      em1977      |      em1978       |      em1979
   |      em1980       |      em1981      |      em1982       |      em1983       |      em1984       |em1985
|      em1986       |      em1987       |      em1988       |      em1989      |      em1990      |      em1991
|      em1992       |      em1993|      em1994      |      em1995       |      em1996       |      em1997
    |      em1998       |      em1999       |      em2000       |      em2001       |em2002       |
em2003      |      em2004       |      em2005      |      em2006       |      em2007       |      em2008
|      em2009      |      em2010       |     gdp1970      |     gdp1971      |     gdp1972      |     gdp1973
|     gdp1974      |     gdp1975      |     gdp1976      |     gdp1977      |gdp1978      |     gdp1979
 |     gdp1980      |     gdp1981      |     gdp1982      |     gdp1983      |     gdp1984      |     gdp1985
      |     gdp1986|    gdp1987      |     gdp1988      |     gdp1989      |     gdp1990      |     gdp1991
     |     gdp1992      |     gdp1993      |     gdp1994      |     gdp1995      |     gdp1996      |
gdp1997      |     gdp1998      |     gdp1999      |     gdp2000      |     gdp2001      |     gdp2002
|     gdp2003      |gdp2004      |     gdp2005      |     gdp2006      |     gdp2007      |     gdp2008
  |     gdp2009      |     gdp2010
-------+---------------------+-------------------+-------------------+------------------+------------------
+------------------+-------------------+------------------+------------------+-------------------+---------
----------+-------------------+------------------+-------------------+-------------------+-----------------
--+------------------+-------------------+-------------------+-------------------+------------------+-------
-----------+------------------+-------------------+-------------------+------------------+------------------
-+-------------------+------------------+-------------------+-------------------+-------------------+-------
------------+--------------------+------------------+-------------------+------------------+----------------
---+-------------------+-------------------+------------------+-------------------+------------------+------
------------+------------------+------------------+------------------+------------------+------------------+
------------------+------------------+------------------+------------------+------------------+-------------
-----+------------------+------------------+------------------+------------------+------------------+-------
-----------+------------------+------------------+------------------+------------------+------------------+-
-----------------+------------------+------------------+------------------+------------------+--------------
----+------------------+------------------+------------------+------------------+------------------+--------
----------+------------------+------------------+------------------+------------------+------------------
0.886 | Belgium             |  18585.6613572407 | -16145.6374560074 |  26938.956253415 | 8094.30475779595 |
 12073.5461203817 | -11069.0567600181 | 19133.8584911727|   5500.312894949 | -4227.94863799987 |  6265.77925410752
|  -10884.749295608 | 30929.4669575201 | -7831.49439429977 |  3235.81760508742 | -22765.9285442662 | 27200
.6767714485 | -10554.9550160917 |   1169.4144482273 | -16783.7961289161 | 27932.2660829329 | 17227.9083196848
| 13956.0524012749 | -40175.6286481088 | -10889.4785920499 | 22703.6576872859 | -14635.5832197402 |
2857.12270512168 | 20473.5044214494 | -52199.4895696423 | -11038.7346460738 |  18466.7298633088 | -17410.4225137703 |
-3475.63826305462 | 29305.6753822341 |   1242.5724942049 | 17491.0096310849 | -12609.9984515902 | -17909.3603476248
|  6276.58431412381 | 21851.9475485178 | -2614.33738160397 | 3777.74134131349 | 4522.08854282736 | 4251.90446379366
| 4512.15101396876 | 4265.49424538129 | 5190.06845330997 | 4543.80444817989 | 5639.81122679089 | 4420.44705213467
|  5658.8820279283 | 5172.69025294376 | 5019.63640408663 | 5938.84979495903 | 4976.57073629812 | 4710.49525137591
| 6523.65700286465 | 5067.82520773578 | 6789.13070219317 | 5525.94643553563 | 6894.68336419297 | 5961.58442474331
| 5661.21093840818 | 7721.56088518218 |  5959.7301109143 | 6453.43604137202 | 6739.39384033096 | 7517.97645468455
| 6907.49136910647 | 7049.03921764209 | 7726.49091035527 | 8552.65909911844 | 7963.94487647115 | 7187.45827585515
| 7994.02955410523 | 9532.89844418041 | 7962.25713582666 | 7846.68238907624 | 10230.9878908643 | 8642.76044946519
| 8886.79860331866 |  8718.3731386891
...
(96 rows)

1.11 - SUMMARIZE_CATCOL

Returns a statistical summary of categorical data input, in three columns:.

Returns a statistical summary of categorical data input, in three columns:

CATEGORY: Categorical levels, of the same SQL data type as the summarized column
COUNT: The number of category levels, of type INTEGER
PERCENT: Represents category percentage, of type FLOAT

Syntax

SUMMARIZE_CATCOL (target-column
        [ USING PARAMETERS TOPK = topk-value [, WITH_TOTALCOUNT = show-total] ] )
OVER()

Arguments

target-column

The name of the input column to summarize, one of the following data types:

BOOLEAN
FLOAT
INTEGER
DATE
CHAR/VARCHAR

Parameters

TOPK: Integer, specifies how many of the most frequent rows to include in the output.
WITH_TOTALCOUNT: A Boolean value that specifies whether the table contains a heading row that displays the total number of rows displayed in the target column, and a percent equal to 100.
Default:true

Examples

This example shows the categorical summary for the current_salary column in the salary_data table. The output of the query shows the column category, count, and percent. The first column gives the categorical levels, with the same SQL data type as the input column, the second column gives a count of that value, and the third column gives a percentage.

=> SELECT SUMMARIZE_CATCOL (current_salary USING PARAMETERS TOPK = 5) OVER() FROM salary_data;
CATEGORY | COUNT | PERCENT
---------+-------+---------
         |  1000 |     100
   39004 |     2 |     0.2
   35321 |     1 |     0.1
   36313 |     1 |     0.1
   36538 |     1 |     0.1
   36562 |     1 |     0.1
(6 rows)

1.12 - SUMMARIZE_NUMCOL

Returns a statistical summary of columns in a Vertica table:.

Returns a statistical summary of columns in a Vertica table:

Count
Mean
Standard deviation
Min/max values
Approximate percentile
Median

All summary values are FLOAT data types, except INTEGER for count.

Syntax

SUMMARIZE_NUMCOL (input-columns [ USING PARAMETERS exclude_columns = 'excluded-columns'] ) OVER()

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. All columns must be a numeric data type. If you select all columns, SUMMARIZE_NUMCOL normalizes all columns in the model

Parameters

exclude_columns: Comma-separated list of column names from input-columns to exclude from processing.

Examples

Show the statistical summary for the age and salary columns in the employee table:

=> SELECT SUMMARIZE_NUMCOL(* USING PARAMETERS exclude_columns='id,name,gender,title') OVER() FROM employee;
COLUMN         | COUNT |    MEAN    |      STDDEV      |  MIN    | PERC25  | MEDIAN  |  PERC75   |  MAX
---------------+-------+------------+------------------+---------+---------+---------+-----------+--------
age            |     5 |    63.4    | 19.3209730603818 |      44 |      45 |      67 |      71   |     90
salary         |     5 | 3456.76    | 1756.78754300285 | 1234.56 | 2345.67 | 3456.78 | 4567.89   | 5678.9
(2 rows)

1.13 - SVD

Computes singular values (the diagonal of the S matrix) and right singular vectors (the V matrix) of an SVD decomposition of the input relation.

Computes singular values (the diagonal of the S matrix) and right singular vectors (the V matrix) of an SVD decomposition of the input relation. The results are saved as an SVD model. The signs of all elements of a singular vector in SVD could be flipped all together on different runs.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

SVD ( 'model-name', 'input-relation', 'input-columns'
     [ USING PARAMETERS
              [exclude_columns = 'excluded-columns']
              [, num_components = num-components]
              [, method = 'method'] ] )

Arguments

model-name: Identifies the model to create, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.
input-relation: The table or view that contains the input data for SVD.
input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. Input columns must be a numeric data type.

Parameters

exclude_columns: Comma-separated list of column names from input-columns to exclude from processing.
num_components: The number of components to keep in the model. The maximum number of components is the number of non-zero singular values computed, which is less than or equal to min (number of columns, number of rows). If you omit this parameter, all components are kept.
method: The method used to calculate SVD, can be set to LAPACK.

Model attributes

columns

The information about columns from the input relation used for creating the SVD model:

index
name

singular_values

The information about singular values found. They are sorted in descending order:

index
value
explained_variance : percentage of the variance in data that can be attributed to this singular value
accumulated_explained_variance : percentage of the variance in data that can be retained if we drop all singular values after this current one

right_singular_vectors

The right singular vectors corresponding to the singular values mentioned above:

index: indices of the elements in each vector
vector1
vector2
...

counters

The information collected during training the model, stored as name-value pairs:

counter_name
- accepted_row_count: number of valid rows in the data
- rejected_row_count: number of invalid rows (having NULL, INF or NaN) in the data
- iteration_count: number of iterations, always 1 for the current implementation of SVD
counter_value

call_string

The function call that created the model.

Privileges

Non-superusers:

CREATE privileges on the schema where the model is created
SELECT privileges on the input relation

Examples

=> SELECT SVD ('svdmodel', 'small_svd', 'x1,x2,x3,x4');
SVD
--------------------------------------------------------------
Finished in 1 iterations.
Accepted Rows: 8  Rejected Rows: 0
(1 row)

=> CREATE TABLE transform_svd AS SELECT
     APPLY_SVD (id, x1, x2, x3, x4 USING PARAMETERS model_name='svdmodel', exclude_columns='id', key_columns='id')
     OVER () FROM small_svd;
CREATE TABLE

=> SELECT * FROM transform_svd;
id  |       col1        |        col2         |        col3         |        col4
----+-------------------+---------------------+---------------------+--------------------
4   |  0.44849499240202 |  -0.347260956311326 |   0.186958376368345 |  0.378561270493651
6   |  0.17652411036246 | -0.0753183783382909 |  -0.678196192333598 | 0.0567124770173372
1   | 0.494871802886819 |   0.161721379259287 |  0.0712816417153664 | -0.473145877877408
2   |  0.17652411036246 | -0.0753183783382909 |  -0.678196192333598 | 0.0567124770173372
3   | 0.150974762654569 |   0.589561842046029 | 0.00392654610109522 |  0.360011163271921
5   | 0.494871802886819 |   0.161721379259287 |  0.0712816417153664 | -0.473145877877408
8   |  0.44849499240202 |  -0.347260956311326 |   0.186958376368345 |  0.378561270493651
7   | 0.150974762654569 |   0.589561842046029 | 0.00392654610109522 |  0.360011163271921
(8 rows)

=> SELECT APPLY_INVERSE_SVD (* USING PARAMETERS model_name='svdmodel', exclude_columns='id',
key_columns='id') OVER () FROM transform_svd;
id  |        x1        |        x2        |        x3        |        x4
----+------------------+------------------+------------------+------------------
4 | 91.4056627665577 | 44.7629617207482 | 83.1704961993117 | 38.9274292265543
6 | 20.6468626294368 | 9.30974906868751 | 8.71006863405534 |  6.5855928603967
7 | 31.2494347777156 | 20.6336519003026 | 27.5668287751507 | 5.84427645886865
1 |  107.93376580719 | 51.6980548011917 | 97.9665796560552 | 40.4918236881051
2 | 20.6468626294368 | 9.30974906868751 | 8.71006863405534 |  6.5855928603967
3 | 31.2494347777156 | 20.6336519003026 | 27.5668287751507 | 5.84427645886865
5 |  107.93376580719 | 51.6980548011917 | 97.9665796560552 | 40.4918236881051
8 | 91.4056627665577 | 44.7629617207482 | 83.1704961993117 | 38.9274292265543
(8 rows)

2 - Machine learning algorithms

Vertica supports a full range of machine learning functions that train a model on a set of data, and return a model that can be saved for later execution.

These functions require the following privileges for non-superusers:

CREATE privileges on the schema where the model is created
SELECT privileges on the input relation

Note

Machine learning algorithms contain a subset of four classification functions:

2.1 - ARIMA

Creates and trains an autoregressive integrated moving average (ARIMA) model from a time series with consistent timesteps.

Creates and trains an autoregressive integrated moving average (ARIMA) model from a time series with consistent timesteps. ARIMA models combine the abilities of AUTOREGRESSOR and MOVING_AVERAGE models by making future predictions based on both preceding time series values and errors of previous predictions. ARIMA models also provide the option to apply a differencing operation to the input data, which can turn a non-stationary time series into a stationary time series. After the model is trained, you can make predictions with the PREDICT_ARIMA function.

In Vertica, ARIMA is implemented using a Kalman Filter state-space approach, similar to Gardner, G., et al. This approach updates the state-space model with each element in the training data in order to calculate a loss score over the training data. A BFGS optimizer is then used to adjust the coefficients, and the state-space estimation is rerun until convergence. Because of this repeated estimation process, ARIMA consumes large amounts of memory when called with high values of p and q.

Given that the input data must be sorted by timestamp, this algorithm is single-threaded.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Immutable

Syntax

ARIMA( 'model-name', 'input-relation', 'timeseries-column', 'timestamp-column'
    USING PARAMETERS param=value[,...] )

Arguments

model-name: Model to create, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.
input-relation: Name of the table or view containing timeseries-column and timestamp-column.
timeseries-column: Name of a NUMERIC column in input-relation that contains the dependent variable or outcome.
timestamp-column: Name of an INTEGER, FLOAT, or TIMESTAMP column in input-relation that represents the timestamp variable. The timestep between consecutive entries should be consistent throughout the timestamp-column.

Tip
If your timestamp-column has varying timesteps, consider standardizing the step size with the TIME_SLICE function.

Parameters

p

Integer in the range [0, 1000], the number of lags to include in the autoregressive component of the computation. If q is unspecified or set to zero, p must be set to a nonzero value. In some cases, using a large p value can result in a memory overload error.

Note

The AUTOREGRESSOR and ARIMA models use different training techniques that produce distinct models when trained with matching parameter values on the same data. For example, if you train an autoregressor model using the same data and p value as an ARIMA model trained with d and q parameters set to zero, those two models will not be identical.

Default: 0

d

Integer in the range [0, 10], the difference order of the model.

If the timeseries-column is a non-stationary time series, whose statistical properties change over time, you can specify a non-zero d value to difference the input data. This operation can remove or reduce trends in the time series data.

Differencing computes the differences between consecutive time series values and then trains the model on these values. The difference order d, where 0 implies no differencing, determines how many times to repeat the differencing operation. For example, second-order differencing takes the results of the first-order operation and differences these values again to obtain the second-order values. For an example that trains an ARIMA model that uses differencing, see ARIMA model example.

Default: 0

q

Integer in the range [0, 1000], the number of lags to include in the moving average component of the computation. If p is unspecified or set to zero, q must be set to a nonzero value. In some cases, using a large q value can result in a memory overload error.

Note

The MOVING_AVERAGE and ARIMA models use different training techniques that produce distinct models when trained with matching parameter values on the same data. For example, if you train a moving-average model using the same data and q value as an ARIMA model trained with p and d parameters set to zero, those two models will not be identical.

Default: 0

missing

Method for handling missing values, one of the following strings:

'drop': Missing values are ignored.
'raise': Missing values raise an error.
'zero': Missing values are set to zero.
'linear_interpolation': Missing values are replaced by a linearly interpolated value based on the nearest valid entries before and after the missing value. In cases where the first or last values in a dataset are missing, the function errors.

Default: 'linear_interpolation'

init_method

Initialization method, one of the following strings:

'Zero': Coefficients are initialized to zero.
'Hannan-Rissanen' or 'HR': Coefficients are initialized using the Hannan-Rissanen algorithm.

Default: 'Zero'

epsilon

Float in the range (0.0, 1.0), controls the convergence criteria of the optimization algorithm.

Default: 1e-6

max_iterations

Integer in the range [1, 1000000), the maximum number of training iterations. If you set this value too low, the algorithm might not converge.

Default: 100

Model attributes

coefficients

Coefficients of the model:

phi: parameters for the autoregressive component of the computation. The number of returned phi values is equal to the value of p.
theta: parameters for the moving average component of the computation. The number of returned theta values is equal to the value of q.

p, q, d

ARIMA component values:

p: number of lags included in the autoregressive component of the computation
d: difference order of the model
q: number of lags included in the moving average component of the computation

mean

The model mean, average of the accepted sample values from timeseries-column

regularization

Type of regularization used when training the model

lambda

Regularization parameter. Higher values indicates stronger regularization.

mean_squared_error

Mean squared error of the model on the training set

rejected_row_count

Number of samples rejected during training

accepted_row_count

Number of samples accepted for training from the data set

timeseries_name

Name of the timeseries-column used to train the model

timestamp_name

Name of the timestamp-column used to train the model

missing_method

Method used for handling missing values

call_string

SQL statement used to train the model

Examples

The function requires that at least one of the p and q parameters be a positive, nonzero integer. The following example trains a model where both of these parameters are set to two:

=> SELECT ARIMA('arima_temp', 'temp_data', 'temperature', 'time' USING PARAMETERS p=2, q=2);
               ARIMA
-------------------------------------
Finished in 24 iterations.
3650 elements accepted, 0 elements rejected.
(1 row)

To see a summary of the model, including all model coefficients and parameter values, call GET_MODEL_SUMMARY:

=> SELECT GET_MODEL_SUMMARY(USING PARAMETERS model_name='arima_temp');
                    GET_MODEL_SUMMARY
------------------------------------------------------------

============
coefficients
============
parameter| value
---------+--------
  phi_1  | 1.23639
  phi_2  |-0.24201
 theta_1 |-0.64535
 theta_2 |-0.23046


==============
regularization
==============
none

===============
timeseries_name
===============
temperature

==============
timestamp_name
==============
time

==============
missing_method
==============
linear_interpolation

===========
call_string
===========
ARIMA('public.arima_temp', 'temp_data', 'temperature', 'time' USING PARAMETERS p=2, d=0, q=2, missing='linear_interpolation', init_method='Zero', epsilon=1e-06, max_iterations=100);

===============
Additional Info
===============
       Name       | Value
------------------+--------
        p         |   2
        q         |   2
        d         |   0
       mean       |11.17775
      lambda      | 1.00000
mean_squared_error| 5.80628
rejected_row_count|   0
accepted_row_count|  3650

(1 row)

For an in-depth example that trains and makes predictions with ARIMA models, see ARIMA model example.

2.2 - AUTOREGRESSOR

Creates an autoregressive (AR) model from a stationary time series with consistent timesteps that can then be used for prediction via PREDICT_AR.

Creates an autoregressive (AR) model from a stationary time series with consistent timesteps that can then be used for prediction via PREDICT_AUTOREGRESSOR.

Autoregressive models predict future values of a time series based on the preceding values. More specifically, the user-specified lag determines how many previous timesteps it takes into account during computation, and predicted values are linear combinations of the values at each lag.

Since its input data must be sorted by timestamp, this algorithm is single-threaded.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

AUTOREGRESSOR ('model-name', 'input-relation', 'data-column', 'timestamp-column'
        [ USING PARAMETERS
              [ p = lags ]
              [, method = 'training-algorithm' ]
              [, missing = 'imputation-method' ]
              [, regularization = 'regularization-method' ]
              [, lambda = regularization-value ]
              [, compute_mse = boolean ]
        ] )

Arguments

model-name: Identifies the model to create, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.
input-relation: The table or view containing the timestamp-column.
This algorithm expects a stationary time series as input; using a time series with a mean that shifts over time may lead to weaker results.
data-column: An input column of type NUMERIC that contains the dependent variables or outcomes.
timestamp-column: One INTEGER, FLOAT, or TIMESTAMP column that represents the timestamp variable. Timesteps must be consistent.

Parameters

p

INTEGER in the range [1, 1999], the number of lags to consider in the computation. Larger values for p weaken the correlation.

Note

Default: 3

method

One of the following algorithms for training the model:

'OLS' (Ordinary Least Squares)
'Yule-Walker'

Note

If you train a model with the Yule-Walker algorithm, the alpha constant in the trained model is set to 0.0.

Default: 'OLS'

missing

One of the following methods for handling missing values:

'drop': Missing values are ignored.
'error': Missing values raise an error.
'zero': Missing values are replaced with 0.
'linear_interpolation': Missing values are replaced by linearly-interpolated values based on the nearest valid entries before and after the missing value. This means that in cases where the first or last values in a dataset are missing, they will simply be dropped.

Default: ' linear_interpolation'

regularization

One of the following regularization methods used when fitting the data:

None
'L2': Weight regularization term which penalizes the squared weight value

Default: None

lambda

FLOAT in the range [0, 100000], the regularization value, lambda.

Default: 1.0

compute_mse

BOOLEAN, whether to calculate and output the mean squared error (MSE).

Default: False

Examples

The following example creates and trains an autoregression model using the Yule-Walker training algorithm and a lag of 3:

=> SELECT AUTOREGRESSOR('AR_temperature_yw', 'temp_data', 'Temperature', 'time' USING PARAMETERS p=3, method='yule-walker');
                   AUTOREGRESSOR
---------------------------------------------------------
Finished. 3650 elements accepted, 0 elements rejected.

(1 row)

See Autoregressive model example for a walk-through of how to train and make predictions with an autoregression model.

2.3 - BISECTING_KMEANS

Executes the bisecting k-means algorithm on an input relation.

Executes the bisecting k-means algorithm on an input relation. The result is a trained model with a hierarchy of cluster centers, with a range of k values, each of which can be used for prediction.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

BISECTING_KMEANS('model-name', 'input-relation', 'input-columns', 'num-clusters'
           [ USING PARAMETERS
                 [exclude_columns = 'exclude-columns']
                 [, bisection_iterations = bisection-iterations]
                 [, split_method = 'split-method']
                 [, min_divisible_cluster_size = min-cluster-size]
                 [, kmeans_max_iterations = kmeans-max-iterations]
                 [, kmeans_epsilon = kmeans-epsilon]
                 [, kmeans_center_init_method = 'kmeans-init-method']
                 [, distance_method = 'distance-method']
                 [, output_view = 'output-view']
                 [, key_columns = 'key-columns'] ] )

Arguments

model-name: Identifies the model to create, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.
input-relation: Table or view that contains the input data for k-means. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.
input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. Input columns must be of data type numeric.
num-clusters: Number of clusters to create, an integer ≤ 10,000. This argument represents the k in k-means.

Parameters

exclude_columns

Comma-separated list of column names from input-columns to exclude from processing.

bisection_iterations

Integer between 1 - 1MM inclusive, specifies number of iterations the bisecting k-means algorithm performs for each bisection step. This corresponds to how many times a standalone k-means algorithm runs in each bisection step.

A setting >1 allows the algorithm to run and choose the best k-means run within each bisection step. If you use kmeanspp, the value of bisection_iterations is always 1, because kmeanspp is more costly to run but also better than the alternatives, so it does not require multiple runs.

Default: 1

split_method

The method used to choose a cluster to bisect/split, one of:

size: Choose the largest cluster to bisect.
sum_squares: Choose the cluster with the largest within-cluster sum of squares to bisect.

Default: sum_squares

min_divisible_cluster_size

Integer ≥ 2, specifies minimum number of points of a divisible cluster.

Default: 2

kmeans_max_iterations

Integer between 1 and 1MM inclusive, specifies the maximum number of iterations the k-means algorithm performs. If you set this value to a number lower than the number of iterations needed for convergence, the algorithm might not converge.

Default: 10

kmeans_epsilon

Integer between 1 and 1MM inclusive, determines whether the k-means algorithm has converged. The algorithm is considered converged after no center has moved more than a distance of epsilon from the previous iteration.

Default: 1e-4

kmeans_center_init_method

The method used to find the initial cluster centers in k-means, one of:

kmeanspp (default): kmeans++ algorithm
pseudo: Uses "pseudo center" approach used by Spark, bisects given center without iterating over points

distance_method

The measure for distance between two data points. Only Euclidean distance is supported at this time.

Default: euclidean

output_view

Name of the view where you save the assignment of each point to its cluster. You must have CREATE privileges on the view schema.

key_columns

Comma-separated list of column names that identify the output rows. Columns must be in the input-columns argument list. To exclude these and other input columns from being used by the algorithm, list them in parameter exclude_columns.

Model attributes

centers

A list of centers of the K centroids.

hierarchy

The hierarchy of K clusters, including:

ParentCluster: Parent cluster centroid of each centroid—that is, the centroid of the cluster from which a cluster is obtained by bisection.
LeftChildCluster: Left child cluster centroid of each centroid—that is, the centroid of the first sub-cluster obtained by bisecting a cluster.
RightChildCluster: the right child cluster centroid of each centroid—that is, the centroid of the second sub-cluster obtained by bisecting a cluster.
BisectionLevel: Specifies which bisection step a cluster is obtained from.
WithinSS: Within-cluster sum of squares for the current cluster
TotalWithinSS: Total within-cluster sum of squares of leaf clusters thus far obtained.

metrics

Several metrics related to the quality of the clustering, including

Total sum of squares
Total within-cluster sum of squares
Between-cluster sum of squares
Between-cluster sum of squares / Total sum of squares
Sum of squares for cluster x, center_id y[...]

Examples

SELECT BISECTING_KMEANS('myModel', 'iris1', '*', '5'
       USING PARAMETERS exclude_columns = 'Species,id', split_method ='sum_squares', output_view = 'myBKmeansView');

2.4 - KMEANS

Executes the k-means algorithm on an input relation.

Executes the k-means algorithm on an input relation. The result is a model with a list of cluster centers.

You can export the resulting k-means model in VERTICA_MODELS or PMML format to apply it on data outside Vertica. You can also train a k-means model elsewhere, then import it to Vertica in PMML format to predict on data in Vertica.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

KMEANS ( 'model-name', 'input-relation', 'input-columns', 'num-clusters'
        [ USING PARAMETERS
           [exclude_columns = 'excluded-columns']
           [, max_iterations = max-iterations]
           [, epsilon = epsilon-value]
           [, { init_method = 'init-method' } | { initial_centers_table = 'init-table' } ]
           [, output_view = 'output-view']
           [, key_columns = 'key-columns'] ] )

Arguments

model-name: Identifies the model to create, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.
input-relation: The table or view that contains the input data for k-means. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.
input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. Input columns must be of data type numeric.
num-clusters: The number of clusters to create, an integer ≤ 10,000. This argument represents the k in k-means.

Parameters

Important

Parameters init_method and initial_centers_table are mutually exclusive. If you set both, the function returns an error.

exclude_columns

Comma-separated list of column names from input-columns to exclude from processing.

max_iterations

The maximum number of iterations the algorithm performs. If you set this value to a number lower than the number of iterations needed for convergence, the algorithm may not converge.

Default: 10

epsilon

Determines whether the algorithm has converged. The algorithm is considered converged after no center has moved more than a distance of
'epsilon' from the previous iteration.

Default: 1e-4

init_method

The method used to find the initial cluster centers, one of the following:

random
kmeanspp (default): kmeans++ algorithm

This value can be memory intensive for high k. If the function returns an error that not enough memory is available, decrease the value of k or use the random method.

initial_centers_table

The table with the initial cluster centers to use. Supply this value if you know the initial centers to use and do not want Vertica to find the initial cluster centers for you.

output_view

The name of the view where you save the assignments of each point to its cluster. You must have CREATE privileges on the schema where the view is saved.

key_columns

Comma-separated list of column names from input-columns that will appear as the columns of output_view. These columns should be picked such that their contents identify each input data point. This parameter is only used if output_view is specified. Columns listed in input-columns that are only meant to be used as key_columns and not for training should be listed in exclude_columns.

Model attributes

centers: A list that contains the center of each cluster.
metrics: A string summary of several metrics related to the quality of the clustering.

Examples

The following example creates k-means model myKmeansModel and applies it to input table iris1. The call to APPLY_KMEANS mixes column names and constants. When a constant is passed in place of a column name, the constant is substituted for the value of the column in all rows:

=> SELECT KMEANS('myKmeansModel', 'iris1', '*', 5
USING PARAMETERS max_iterations=20, output_view='myKmeansView', key_columns='id', exclude_columns='Species, id');
           KMEANS
----------------------------
 Finished in 12 iterations

(1 row)
=> SELECT id, APPLY_KMEANS(Sepal_Length, 2.2, 1.3, Petal_Width
USING PARAMETERS model_name='myKmeansModel', match_by_pos='true') FROM iris2;
 id  | APPLY_KMEANS
-----+--------------
   5 |            1
  10 |            1
  14 |            1
  15 |            1
  21 |            1
  22 |            1
  24 |            1
  25 |            1
  32 |            1
  33 |            1
  34 |            1
  35 |            1
  38 |            1
  39 |            1
  42 |            1
...
 (60 rows)

2.5 - KPROTOTYPES

Executes the k-prototypes algorithm on an input relation.

Executes the k-prototypes algorithm on an input relation. The result is a model with a list of cluster centers.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Syntax

SELECT KPROTOTYPES ('`*`model-name`*`', '`*`input-relation`*`', '`*`input-columns`*`', `*`num-clusters`*`
                [USING PARAMETERS [exclude_columns = '`*`exclude-columns`*`']
                [, max_iterations = '`*`max-iterations`*`']
                [, epsilon = `*`epsilon`*`]
                [, {[init_method = '`*`init-method`*`'] } | { initial_centers_table = '`*`init-table`*`' } ]
                [, gamma = '`*`gamma`*`']
                [, output_view = '`*`output-view`*`']
                [, key_columns = '`*`key-columns`*`']]);

Behavior type

Volatile

Arguments

model-name: Name of the model resulting from the training.
input-relation: Name of the table or view containing the training samples.
input-columns: String containing a comma-separated list of columns to use from the input-relation, or asterisk (*) to select all columns.
num-clusters: Integer ≤ 10,000 representing the number of clusters to create. This argument represents the k in k-prototypes.

Parameters

exclude-columns: String containing a comma-separated list of column names from input-columns to exclude from processing.
Default: (empty)
max_iterations: Integer ≤ 1M representing the maximum number of iterations the algorithm performs.
Default: Integer ≤ 1M
epsilon: Integer which determines whether the algorithm has converged.
Default: 1e-4
init_method: String specifying the method used to find the initial k-prototypes cluster centers.
Default: "random"
initial_centers_table: The table with the initial cluster centers to use.
gamma: Float between 0 and 10000 specifying the weighing factor for categorical columns. It can determine relative importance of numerical and categorical attributes
Default: Inferred from data.
output_view: The name of the view where you save the assignments of each point to its cluster
key_columns: Comma-separated list of column names that identify the output rows. Columns must be in the input-columns argument list

Examples

The following example creates k-prototypes model small_model and applies it to input table small_test_mixed:

=> SELECT KPROTOTYPES('small_model_initcenters', 'small_test_mixed', 'x0, country', 3 USING PARAMETERS initial_centers_table='small_test_mixed_centers', key_columns='pid');
      KPROTOTYPES
---------------------------
Finished in 2 iterations

(1 row)

=> SELECT country, x0, APPLY_KPROTOTYPES(country, x0
USING PARAMETERS model_name='small_model')
FROM small_test_mixed;
  country   | x0  | apply_kprototypes
------------+-----+-------------------
 'China'    |  20 |                 0
 'US'       |  85 |                 2
 'Russia'   |  80 |                 1
 'Brazil'   |  78 |                 1
 'US'       |  23 |                 0
 'US'       |  50 |                 0
 'Canada'   |  24 |                 0
 'Canada'   |  18 |                 0
 'Russia'   |  90 |                 2
 'Russia'   |  98 |                 2
 'Brazil'   |  89 |                 2
...
(45 rows)

2.6 - LINEAR_REG

Executes linear regression on an input relation, and returns a linear regression model.

You can export the resulting linear regression model in VERTICA_MODELS or PMML format to apply it on data outside Vertica. You can also train a linear regression model elsewhere, then import it to Vertica in PMML format to model on data inside Vertica.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

LINEAR_REG ( 'model-name', 'input-relation', 'response-column', 'predictor-columns'
        [ USING PARAMETERS
              [exclude_columns = 'excluded-columns']
              [, optimizer = 'optimizer-method']
              [, regularization = 'regularization-method']
              [, epsilon = epsilon-value]
              [, max_iterations = iterations]
              [, lambda = lamda-value]
              [, alpha = alpha-value]
              [, fit_intercept = boolean-value] ] )

Arguments

model-name

input-relation

Table or view that contains the training data for building the model. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.

response-column

Name of the input column that represents the dependent variable or outcome. All values in this column must be numeric, otherwise the model is invalid.

predictor-columns

Comma-separated list of columns in the input relation that represent independent variables for the model, or asterisk (*) to select all columns. If you select all columns, the argument list for parameter exclude_columns must include response-column, and any columns that are invalid as predictor columns.

All predictor columns must be of type numeric or BOOLEAN; otherwise the model is invalid.

Note

All BOOLEAN predictor values are converted to FLOAT values before training: 0 for false, 1 for true. No type checking occurs during prediction, so you can use a BOOLEAN predictor column in training, and during prediction provide a FLOAT column of the same name. In this case, all FLOAT values must be either 0 or 1.

Parameters

exclude_columns

Comma-separated list of columns from predictor-columns to exclude from processing.

optimizer

Optimizer method used to train the model, one of the following:

Newton
BFGS
CGD

Important
If you select CGD, regularization-method must be set to L1 or ENet, otherwise the function returns an error.

Default: CGD if regularization-method is set to L1 or ENet, otherwise Newton.

regularization

Method of regularization, one of the following:

None (default)
L1
L2
ENet

epsilon

FLOAT in the range (0.0, 1.0), the error value at which to stop training. Training stops if either the difference between the actual and predicted values is less than or equal to epsilon or if the number of iterations exceeds max_iterations.

Default: 1e-6

max_iterations

INTEGER in the range (0, 1000000), the maximum number of training iterations. Training stops if either the number of iterations exceeds max_iterations or if the difference between the actual and predicted values is less than or equal to epsilon.

Default: 100

lambda

Integer ≥ 0, specifies the value of the regularization parameter.

Default: 1

alpha

Integer ≥ 0, specifies the value of the ENET regularization parameter, which defines how much L1 versus L2 regularization to provide. A value of 1 is equivalent to L1 and a value of 0 is equivalent to L2.

Value range: [0,1]

Default: 0.5

fit_intercept

Boolean, specifies whether the model includes an intercept. By setting to false, no intercept will be used in training the model. Note that setting fit_intercept to false does not work well with the BFGS optimizer.

Default: True

Model attributes

data

The data for the function, including:

coeffNames: Name of the coefficients. This starts with intercept and then follows with the names of the predictors in the same order specified in the call.
coeff: Vector of estimated coefficients, with the same order as coeffNames
stdErr: Vector of the standard error of the coefficients, with the same order as coeffNames
zValue (for logistic regression): Vector of z-values of the coefficients, in the same order as coeffNames
tValue (for linear regression): Vector of t-values of the coefficients, in the same order as coeffNames
pValue: Vector of p-values of the coefficients, in the same order as coeffNames

regularization

Type of regularization to use when training the model.

lambda

Regularization parameter. Higher values enforce stronger regularization. This value must be nonnegative.

alpha

Elastic net mixture parameter.

iterations

Number of iterations that actually occur for the convergence before exceeding max_iterations.

skippedRows

Number of rows of the input relation that were skipped because they contained an invalid value.

processedRows

Total number of input relation rows minus skippedRows.

callStr

Value of all input arguments specified when the function was called.

Examples

=> SELECT LINEAR_REG('myLinearRegModel', 'faithful', 'eruptions', 'waiting'
                      USING PARAMETERS optimizer='BFGS', fit_intercept=true);
         LINEAR_REG
----------------------------
 Finished in 10 iterations

(1 row)

2.7 - LOGISTIC_REG

Executes logistic regression on an input relation.

Executes logistic regression on an input relation. The result is a logistic regression model.

You can export the resulting logistic regression model in VERTICA_MODELS or PMML format to apply it on data outside Vertica. You can also train a logistic regression model elsewhere, then import it to Vertica in PMML format to predict on data in Vertica.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

LOGISTIC_REG ( 'model-name', 'input-relation', 'response-column', 'predictor-columns'
        [ USING PARAMETERS [exclude_columns = 'excluded-columns']
              [, optimizer = 'optimizer-method']
              [, regularization = 'regularization-method']
              [, epsilon = epsilon-value]
              [, max_iterations = iterations]
              [, lambda = lamda-value]
              [, alpha = alpha-value]
              [, fit_intercept = boolean-value] ] )

Arguments

model-name

input-relation

The table or view that contains the training data for building the model. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.

response-column

The input column that represents the dependent variable or outcome. The column value must be 0 or 1, and of type numeric or BOOLEAN. The function automatically skips all other values.

predictor-columns

All predictor columns must be of type numeric or BOOLEAN; otherwise the model is invalid.

Note

Parameters

exclude_columns

Comma-separated list of columns from predictor-columns to exclude from processing.

optimizer

The optimizer method used to train the model, one of the following:

Newton
BFGS
CGD

Important
If you select CGD, regularization-method must be set to L1 or ENet, otherwise the function returns an error.

Default: CGD if regularization-method is set to L1 or ENet, otherwise Newton.

regularization

The method of regularization, one of the following:

None (default)
L1
L2
ENet

epsilon

Default: 1e-6

max_iterations

Default: 100

lambda

Integer ≥ 0, specifies the value of the regularization parameter.

Default: 1

alpha

Value range: [0,1]

Default: 0.5

fit_intercept

Default: True

Model attributes

data

The data for the function, including:

coeffNames: Name of the coefficients. This starts with intercept and then follows with the names of the predictors in the same order specified in the call.
coeff: Vector of estimated coefficients, with the same order as coeffNames
stdErr: Vector of the standard error of the coefficients, with the same order as coeffNames
zValue (for logistic regression): Vector of z-values of the coefficients, in the same order as coeffNames
tValue (for linear regression): Vector of t-values of the coefficients, in the same order as coeffNames
pValue: Vector of p-values of the coefficients, in the same order as coeffNames

regularization

Type of regularization to use when training the model.

lambda

Regularization parameter. Higher values enforce stronger regularization. This value must be nonnegative.

alpha

Elastic net mixture parameter.

iterations

Number of iterations that actually occur for the convergence before exceeding max_iterations.

skippedRows

Number of rows of the input relation that were skipped because they contained an invalid value.

processedRows

Total number of input relation rows minus skippedRows.

callStr

Value of all input arguments specified when the function was called.

Privileges

Superuser, or SELECT privileges on the input relation

Examples

=> SELECT LOGISTIC_REG('myLogisticRegModel', 'mtcars', 'am',
                       'mpg, cyl, disp, hp, drat, wt, qsec, vs, gear, carb'
                        USING PARAMETERS exclude_columns='hp', optimizer='BFGS', fit_intercept=true);
        LOGISTIC_REG
----------------------------
 Finished in 20 iterations

(1 row)

2.8 - MOVING_AVERAGE

Creates a moving-average (MA) model from a stationary time series with consistent timesteps that can then be used for prediction via PREDICT_MOVING_AVERAGE.

Moving average models use the errors of previous predictions to make future predictions. More specifically, the user-specified lag determines how many previous predictions and errors it takes into account during computation.

Since its input data must be sorted by timestamp, this algorithm is single-threaded.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

MOVING_AVERAGE ('model-name', 'input-relation', 'data-column', 'timestamp-column'
        [ USING PARAMETERS
              [ q = lags ]
              [, missing = "imputation-method" ]
              [, regularization = "regularization-method" ]
              [, lambda = regularization-value ]
              [, compute_mse = boolean ]
        ] )

Arguments

model-name: Identifies the model to create, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.
input-relation: The table or view containing the timestamp-column.
This algorithm expects a stationary time series as input; using a time series with a mean that shifts over time may lead to weaker results.
data-column: An input column of type NUMERIC that contains the dependent variables or outcomes.
timestamp-column: One INTEGER, FLOAT, or TIMESTAMP column that represent the timestamp variable. Timesteps must be consistent.

Parameters

q

INTEGER in the range [1, 67), the number of lags to consider in the computation.

Note

Default: 1

missing

One of the following methods for handling missing values:

drop: Missing values are ignored.
error: Missing values raise an error.
zero: Missing values are replaced with 0.
linear_interpolation: Missing values are replaced by linearly interpolated values based on the nearest valid entries before and after the missing value. This means that in cases where the first or last values in a dataset are missing, they will simply be dropped.

Default: linear_interpolation

regularization

One of the following regularization methods used when fitting the data:

None
L2: weight regularization term which penalizes the squared weight value

Default: None

lambda

FLOAT in the range [0, 100000], the regularization value, lambda.

Default: 1.0

compute_mse

BOOLEAN, whether to calculate and output the mean squared error (MSE).

This parameter only accepts "true" or "false" rather than the standard literal equivalents for BOOLEANs like 1 or 0.

Default: False

Examples

See Moving-average model example.

2.9 - NAIVE_BAYES

Executes the Naive Bayes algorithm on an input relation and returns a Naive Bayes model.

Columns are treated according to data type:

FLOAT: Values are assumed to follow some Gaussian distribution.
INTEGER: Values are assumed to belong to one multinomial distribution.
CHAR/VARCHAR: Values are assumed to follow some categorical distribution. The string values stored in these columns must not be greater than 128 characters.
BOOLEAN: Values are treated as categorical with two values.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

NAIVE_BAYES ( 'model-name', 'input-relation', 'response-column', 'predictor-columns'
        [ USING PARAMETERS [exclude_columns = 'excluded-columns'] [, alpha = alpha-value] ] )

Arguments

model-name

input-relation

response-column

Name of the input column that represents the dependent variable, or outcome. This column must contain discrete labels that represent different class labels.

The response column must be of type numeric, CHAR/VARCHAR, or BOOLEAN; otherwise the model is invalid.

Note

Vertica automatically casts numeric response column values to VARCHAR.

predictor-columns

All predictor columns must be of type numeric, CHAR/VARCHAR, or BOOLEAN; otherwise the model is invalid. BOOLEAN column values are converted to FLOAT values before training: 0 for false, 1 for true.

Parameters

exclude_columns: Comma-separated list of columns from predictor-columns to exclude from processing.
alpha: Float, specifies use of Laplace smoothing if the event model is categorical, multinomial, or Bernoulli.
Default: 1.0

Model attributes

colsInfo

The information from the response and predictor columns used in training:

index: The index (starting at 0) of the column as provided in training. Index 0 is used for the response column.
name: The column name.
type: The label used for response with a value of Gaussian, Multinominal, Categorical, or Bernoulli.

alpha

The smooth parameter value.

prior

The percentage of each class among all training samples:

label: The class label.
value: The percentage of each class.

nRowsTotal

The number of samples accepted for training from the data set.

nRowsRejected

The number of samples rejected for training.

callStr

The SQL statement used to replicate the training.

Gaussian

The Gaussian model conditioned on the class indicated by the class_name:

index: The index of the predictor column.
mu: The mean value of the model.
sigmaSq: The squared standard deviation of the model.

Multinominal

The Multinomial model conditioned on the class indicated by the class_name:

index: The index of the predictor column.
prob: The probability conditioned on the class indicated by the class_name.

Bernoulli

The Bernoulli model conditioned on the class indicated by the class_name:

index: The index of the predictor column.
probTrue: The probability of having the value TRUE in this predictor column.

Categorical

The Gaussian model conditioned on the class indicated by the class_name:

category: The value in the predictor name.
<class_name>: The probability of having that value conditioned on the class indicated by the class_name.

Privileges

Superuser, or SELECT privileges on the input relation.

Examples

=> SELECT NAIVE_BAYES('naive_house84_model', 'house84_train', 'party', '*'
                      USING PARAMETERS exclude_columns='party, id');
                                  NAIVE_BAYES
--------------------------------------------------
 Finished. Accepted Rows: 324  Rejected Rows: 0
(1 row)

2.10 - POISSON_REG

Executes Poisson regression on an input relation, and returns a Poisson regression model.

You can export the resulting Poisson regression model in VERTICA_MODELS or PMML format to apply it on data outside Vertica. You can also train a Poisson regression model elsewhere, then import it to Vertica in PMML format to apply it on data inside Vertica.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

POISSON_REG ( 'model-name', 'input-table', 'response-column', 'predictor-columns'
        [ USING PARAMETERS
              [exclude_columns = 'excluded-columns']
              [, optimizer = 'optimizer-method']
              [, regularization = 'regularization-method']
              [, epsilon = epsilon-value]
              [, max_iterations = iterations]
              [, lambda = lamda-value]
              [, fit_intercept = boolean-value] ] )

Arguments

model-name

input-table

response-column

Name of input column that represents the dependent variable or outcome. All values in this column must be numeric, otherwise the model is invalid.

predictor-columns

All predictor columns must be of type numeric or BOOLEAN; otherwise the model is invalid.

Note

Parameters

exclude_columns

Comma-separated list of columns from predictor-columns to exclude from processing.

optimizer

Optimizer method used to train the model. The currently supported method is Newton.

regularization

Method of regularization, one of the following:

None (default)
L2

epsilon

FLOAT in the range (0.0, 1.0), the error value at which to stop training. Training stops if either the relative change in Poisson deviance is less than or equal to epsilon or if the number of iterations exceeds max_iterations.

Default: 1e-6

max_iterations

INTEGER in the range (0, 1000000), the maximum number of training iterations. Training stops if either the number of iterations exceeds max_iterations or the relative change in Poisson deviance is less than or equal to epsilon.

lambda

FLOAT ≥ 0, specifies the regularization strength.

Default: 1.0

fit_intercept

Boolean, specifies whether the model includes an intercept. By setting to false, no intercept will be used in training the model.”

Default: True

Model attributes

data

Data for the function, including:

coeffNames: Name of the coefficients. This starts with intercept and then follows with the names of the predictors in the same order specified in the call.
coeff: Vector of estimated coefficients, with the same order as coeffNames
stdErr: Vector of the standard error of the coefficients, with the same order as coeffNames
zValue: (for logistic and Poisson regression): Vector of z-values of the coefficients, in the same order as coeffNames
tValue (for linear regression): Vector of t-values of the coefficients, in the same order as coeffNames
pValue: Vector of p-values of the coefficients, in the same order as coeffNames

regularization

Type of regularization to use when training the model.

lambda

Regularization parameter. Higher values enforce stronger regularization. This value must be nonnegative.

iterations

Number of iterations that actually occur for the convergence before exceeding max_iterations.

skippedRows

Number of rows of the input relation that were skipped because they contained an invalid value.

processedRows

Total number of input relation rows minus skippedRows.

callStr

Value of all input arguments specified when the function was called.

Examples

=> SELECT POISSON_REG('myModel', 'numericFaithful', 'eruptions', 'waiting' USING PARAMETERS epsilon=1e-8);
poisson_reg
---------------------------
Finished in 7 iterations

(1 row)

2.11 - RF_CLASSIFIER

Trains a random forest model for classification on an input relation.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

RF_CLASSIFIER ( 'model-name', input-relation, 'response-column', 'predictor-columns'
        [ USING PARAMETERS
              [exclude_columns = 'excluded-columns']
              [, ntree = num-trees]
              [, mtry = num-features]
              [, sampling_size = sampling-size]
              [, max_depth = depth]
              [, max_breadth = breadth]
              [, min_leaf_size = leaf-size]
              [, min_info_gain = threshold]
              [, nbins = num-bins] ] )

Arguments

model-name

Identifies the model stored as a result of the training, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.

input-relation

The table or view that contains the training samples. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.

response-column

An input column of type numeric, CHAR/VARCHAR, or BOOLEAN that represents the dependent variable.

Note

Vertica automatically casts numeric response column values to VARCHAR.

predictor-columns

All predictor columns must be of type numeric, CHAR/VARCHAR, or BOOLEAN; otherwise the model is invalid.

Vertica XGBoost and Random Forest algorithms offer native support for categorical columns (BOOL/VARCHAR). Simply pass the categorical columns as predictors to the models and the algorithm will automatically treat the columns as categorical and will not attempt to split them into bins in the same manner as numerical columns; Vertica treats these columns as true categorical values and does not simply cast them to continuous values under-the-hood.

Parameters

exclude_columns

Comma-separated list of column names from input-columns to exclude from processing.

ntree

Integer in the range [1, 1000], the number of trees in the forest.

Default: 20

mtry

Integer in the range [1, number-predictors], the number of randomly chosen features from which to pick the best feature to split on a given tree node.

Default: Square root of the total number of predictors

sampling_size

Float in the range (0.0, 1.0], the portion of the input data set that is randomly picked for training each tree.

Default: 0.632

max_depth

Integer in the range [1, 100], the maximum depth for growing each tree. For example, a max_depth of 0 represents a tree with only a root node, and a max_depth of 2 represents a tree with four leaf nodes.

Default: 5

max_breadth

Integer in the range [1, 1e9], the maximum number of leaf nodes a tree can have.

Default: 32

min_leaf_size

Integer in the range [1, 1e6], the minimum number of samples each branch must have after splitting a node. A split that results in fewer remaining samples in its left or right branch is be discarded, and the node is treated as a leaf node.

Default: 1

min_info_gain

Float in the range [0.0, 1.0), the minimum threshold for including a split. A split with information gain less than this threshold is discarded.

Default: 0.0

nbins

Integer in the range [2, 1000], the number of bins to use for discretizing continuous features.

Default: 32

Model attributes

data

Data for the function, including:

predictorNames: The name of the predictors in the same order they were specified for training the model.
predictorTypes: The type of the predictors in the same order as their names in predictorNames.

ntree

Number of trees in the model.

skippedRows

Number of rows in input_relation that were skipped because they contained an invalid value.

processedRows

Total number of rows in input_relation minus skippedRows.

callStr

Value of all input arguments that were specified at the time the function was called.

Examples

=> SELECT RF_CLASSIFIER ('myRFModel', 'iris', 'Species', 'Sepal_Length, Sepal_Width,
Petal_Length, Petal_Width' USING PARAMETERS ntree=100, sampling_size=0.3);

RF_CLASSIFIER
--------------------------------------------------
Finished training
(1 row)

2.12 - RF_REGRESSOR

Trains a random forest model for regression on an input relation.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

RF_REGRESSOR ( 'model-name', input-relation, 'response-column', 'predictor-columns'
        [ USING PARAMETERS
              [exclude_columns = 'excluded-columns']
              [, ntree = num-trees]
              [, mtry = num-features]
              [, sampling_size = sampling-size]
              [, max_depth = depth]
              [, max_breadth = breadth]
              [, min_leaf_size = leaf-size]
              [, min_info_gain = threshold]
              [, nbins = num-bins] ] )

Arguments

model-name

The model that is stored as a result of training, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.

input-relation

response-column

A numeric input column that represents the dependent variable.

predictor-columns

All predictor columns must be of type numeric, CHAR/VARCHAR, or BOOLEAN; otherwise the model is invalid.

Parameters

exclude_columns

Comma-separated list of columns from predictor-columns to exclude from processing.

ntree

Integer in the range [1, 1000], the number of trees in the forest.

Default: 20

mtry

Integer in the range [1, number-predictors], the number of features to consider at the split of a tree node.

Default: One-third the total number of predictors

sampling_size

Float in the range (0.0, 1.0], the portion of the input data set that is randomly picked for training each tree.

Default: 0.632

max_depth

Default: 5

max_breadth

Integer in the range [1, 1e9], the maximum number of leaf nodes a tree can have.

Default: 32

min_leaf_size

The default value of this parameter differs from that of analogous parameters in libraries like sklearn and will therefore yield a model with predicted values that differ from the original response values.

Default: 5

min_info_gain

Float in the range [0.0, 1.0), the minimum threshold for including a split. A split with information gain less than this threshold is discarded.

Default: 0.0

nbins

Integer in the range [2, 1000], the number of bins to use for discretizing continuous features.

Default: 32

Model attributes

data

Data for the function, including:

predictorNames: The name of the predictors in the same order they were specified for training the model.
predictorTypes: The type of the predictors in the same order as their names in predictorNames.

ntree

Number of trees in the model.

skippedRows

Number of rows in input_relation that were skipped because they contained an invalid value.

processedRows

Total number of rows in input_relation minus skippedRows.

callStr

Value of all input arguments that were specified at the time the function was called.

Examples

=> SELECT RF_REGRESSOR ('myRFRegressorModel', 'mtcars', 'carb', 'mpg, cyl, hp, drat, wt' USING PARAMETERS
ntree=100, sampling_size=0.3);
RF_REGRESSOR
--------------
Finished
(1 row)

2.13 - SVM_CLASSIFIER

Trains the SVM model on an input relation.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

SVM_CLASSIFIER ( 'model-name', input-relation, 'response-column', 'predictor-columns'
        [ USING PARAMETERS
              [exclude_columns = 'excluded-columns']
              [, C = 'cost']
              [, epsilon = 'epsilon-value']
              [, max_iterations = 'max-iterations']
              [, class_weights = 'weight']
              [, intercept_mode = 'intercept-mode']
              [, intercept_scaling = 'scale'] ] )

Arguments

model-name

input-relation

The table or view that contains the training data. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.

response-column

The input column that represents the dependent variable or outcome. The column value must be 0 or 1, and of type numeric or BOOLEAN, otherwise the function returns with an error.

predictor-columns

All predictor columns must be of type numeric or BOOLEAN; otherwise the model is invalid.

Note

Parameters

exclude_columns

Comma-separated list of columns from predictor-columns to exclude from processing.

C

Weight for misclassification cost. The algorithm minimizes the regularization cost and the misclassification cost.

Default: 1.0

epsilon

Used to control accuracy.

Default: 1e-3

max_iterations

Maximum number of iterations that the algorithm performs.

Default: 100

class_weights

Specifies how to determine weights of the two classes, one of the following:

None (default): No weights are used
value0, value1: Two comma-delimited strings that specify two positive FLOAT values, where value0 assigns a weight to class 0, and value1 assigns a weight to class 1.
auto: Weights each class according to the number of samples.

intercept_mode

Specifies how to treat the intercept, one of the following:

regularized (default): Fits the intercept and applies a regularization on it.
unregularized: Fits the intercept but does not include it in regularization.

intercept_scaling

Float value that serves as the value of a dummy feature whose coefficient Vertica uses to calculate the model intercept. Because the dummy feature is not in the training data, its values are set to a constant, by default 1.

Model attributes

coeff

Coefficients in the model:

colNames: Intercept, or predictor column name
coefficients: Coefficient value

nAccepted

Number of samples accepted for training from the data set

nRejected

Number of samples rejected when training

nIteration

Number of iterations used in training

callStr

SQL statement used to replicate the training

Examples

The following example uses SVM_CLASSIFIER on the mtcars table:


=> SELECT SVM_CLASSIFIER(
       'mySvmClassModel', 'mtcars', 'am', 'mpg,cyl,disp,hp,drat,wt,qsec,vs,gear,carb'
       USING PARAMETERS exclude_columns = 'hp,drat');
SVM_CLASSIFIER
----------------------------------------------------------------
Finished in 15 iterations.
Accepted Rows: 32  Rejected Rows: 0
(1 row)

2.14 - SVM_REGRESSOR

Trains the SVM model on an input relation.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

SVM_REGRESSOR ( 'model-name', input-relation, 'response-column', 'predictor-columns'
        [ USING PARAMETERS
              [exclude_columns = 'excluded-columns']
              [, error_tolerance = error-tolerance]
              [, C = cost]
              [, epsilon = epsilon-value]
              [, max_iterations = max-iterations]
              [, intercept_mode = 'mode']
              [, intercept_scaling = 'scale'] ] )

Arguments

model-name

input-relation

The table or view that contains the training data. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.

response-column

An input column that represents the dependent variable or outcome. The column must be a numeric data type.

predictor-columns

All predictor columns must be of type numeric or BOOLEAN; otherwise the model is invalid.

Note

Parameters

exclude_columns

Comma-separated list of columns from predictor-columns to exclude from processing.

error_tolerance

Defines the acceptable error margin. Any data points outside this region add a penalty to the cost function.

Default: 0.1

C

The weight for misclassification cost. The algorithm minimizes the regularization cost and the misclassification cost.

Default: 1.0

epsilon

Used to control accuracy.

Default: 1e-3

max_iterations

The maximum number of iterations that the algorithm performs.

Default: 100

intercept_mode

A string that specifies how to treat the intercept, one of the following

regularized (default): Fits the intercept and applies a regularization on it.
unregularized: Fits the intercept but does not include it in regularization.

intercept_scaling

A FLOAT value, serves as the value of a dummy feature whose coefficient Vertica uses to calculate the model intercept. Because the dummy feature is not in the training data, its values are set to a constant, by default set to 1.

Model attributes

coeff

Coefficients in the model:

colNames: Intercept, or predictor column name
coefficients: Coefficient value

nAccepted

Number of samples accepted for training from the data set

nRejected

Number of samples rejected when training

nIteration

Number of iterations used in training

callStr

SQL statement used to replicate the training

Examples


=> SELECT SVM_REGRESSOR('mySvmRegModel', 'faithful', 'eruptions', 'waiting'
                          USING PARAMETERS error_tolerance=0.1, max_iterations=100);
SVM_REGRESSOR
----------------------------------------------------------------
Finished in 5 iterations.
Accepted Rows: 272  Rejected Rows: 0
(1 row)

2.15 - XGB_CLASSIFIER

Trains an XGBoost model for classification on an input relation.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

XGB_CLASSIFIER ('model-name', 'input-relation', 'response-column', 'predictor-columns'
        [ USING PARAMETERS param=value[,...] ] )

Arguments

model-name

Name of the model (case-insensitive).

input-relation

response-column

An input column of type CHAR or VARCHAR that represents the dependent variable or outcome.

predictor-columns

Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. Columns must be of data types CHAR, VARCHAR, BOOL, INT, or FLOAT.

Columns of type CHAR, VARCHAR, and BOOL are treated as categorical features; all others are treated as numeric features.

Parameters

exclude_columns

Comma-separated list of column names from input-columns to exclude from processing.

max_ntree

Integer in the range [1,1000] that sets the maximum number of trees to create.

Default: 10

max_depth

Integer in the range [1,40] that specifies the maximum depth of each tree.

Default: 6

objective

The objective/loss function used to iteratively improve the model. 'crossentropy' is the only option.

Default: 'crossentropy'

split_proposal_method

The splitting strategy for the feature columns. 'global' is the only option. This method calculates the split for each feature column only at the beginning of the algorithm. The feature columns are split into the number of bins specified by nbins.

Default: 'global'

learning_rate

Float in the range (0,1] that specifies the weight for each tree's prediction. Setting this parameter can reduce each tree's impact and thereby prevent earlier trees from monopolizing improvements at the expense of contributions from later trees.

Default: 0.3

min_split_loss

Float in the range [0,1000] that specifies the minimum amount of improvement each split must achieve on the model's objective function value to avoid being pruned.

If set to 0 or omitted, no minimum is set. In this case, trees are pruned according to positive or negative objective function values.

Default: 0.0 (disable)

weight_reg

Float in the range [0,1000] that specifies the regularization term applied to the weights of classification tree leaves. The higher the setting, the sparser or smoother the weights are, which can help prevent over-fitting.

Default: 1.0

nbins

Integer in the range (1,1000] that specifies the number of bins to use for finding splits in each column. More bins leads to longer runtime but more fine-grained and possibly better splits.

Default: 32

sampling_size

Float in the range (0,1] that specifies the fraction of rows to use in each training iteration.

A value of 1 indicates that all rows are used.

Default: 1.0

col_sample_by_tree

Float in the range (0,1] that specifies the fraction of columns (features), chosen at random, to use when building each tree.

A value of 1 indicates that all columns are used.

col_sample_by parameters "stack" on top of each other if several are specified. That is, given a set of 24 columns, for col_sample_by_tree=0.5 andcol_sample_by_node=0.5,col_sample_by_tree samples 12 columns, reducing the available, unsampled column pool to 12. col_sample_by_node then samples half of the remaining pool, so each node samples 6 columns.

This algorithm will always sample at least one column.

Default: 1

col_sample_by_node

Float in the range (0,1] that specifies the fraction of columns (features), chosen at random, to use when evaluating each split.

A value of 1 indicates that all columns are used.

This algorithm will always sample at least one column.

Default: 1

Examples

See XGBoost for classification.

2.16 - XGB_REGRESSOR

Trains an XGBoost model for regression on an input relation.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

XGB_REGRESSOR ('model-name', 'input-relation', 'response-column', 'predictor-columns'
        [ USING PARAMETERS param=value[,...] ] )

Arguments

model-name

Name of the model (case-insensitive).

input-relation

response-column

An input column of type INTEGER or FLOAT that represents the dependent variable or outcome.

predictor-columns

Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. Columns must be of data types CHAR, VARCHAR, BOOL, INT, or FLOAT.

Columns of type CHAR, VARCHAR, and BOOL are treated as categorical features; all others are treated as numeric features.

Parameters

exclude_columns

Comma-separated list of column names from input-columns to exclude from processing.

max_ntree

Integer in the range [1,1000] that sets the maximum number of trees to create.

Default: 10

max_depth

Integer in the range [1,40] that specifies the maximum depth of each tree.

Default: 6

objective

The objective/loss function used to iteratively improve the model. 'squarederror' is the only option.

Default: 'squarederror'

split_proposal_method

Default: 'global'

learning_rate

Default: 0.3

min_split_loss

Float in the range [0,1000] that specifies the minimum amount of improvement each split must achieve on the model's objective function value to avoid being pruned.

If set to 0 or omitted, no minimum is set. In this case, trees are pruned according to positive or negative objective function values.

Default: 0.0 (disable)

weight_reg

Default: 1.0

nbins

Integer in the range (1,1000] that specifies the number of bins to use for finding splits in each column. More bins leads to longer runtime but more fine-grained and possibly better splits.

Default: 32

sampling_size

Float in the range (0,1] that specifies the fraction of rows to use in each training iteration.

A value of 1 indicates that all rows are used.

Default: 1.0

col_sample_by_tree

Float in the range (0,1] that specifies the fraction of columns (features), chosen at random, to use when building each tree.

A value of 1 indicates that all columns are used.

This algorithm will always sample at least one column.

Default: 1

col_sample_by_node

Float in the range (0,1] that specifies the fraction of columns (features), chosen at random, to use when evaluating each split.

A value of 1 indicates that all columns are used.

This algorithm will always sample at least one column.

Default: 1

Examples

See XGBoost for regression.

3 - Model evaluation

A set of Vertica machine learning functions evaluate the prediction data that is generated by trained models, or return information about the models themselves.

3.1 - CONFUSION_MATRIX

Computes the confusion matrix of a table with observed and predicted values of a response variable.

Computes the confusion matrix of a table with observed and predicted values of a response variable. CONFUSION_MATRIX produces a table with the following dimensions:

Rows: Number of classes
Columns: Number of classes + 2

Syntax

CONFUSION_MATRIX ( targets, predictions [ USING PARAMETERS num_classes = num-classes ] OVER()

Arguments

targets: An input column that contains the true values of the response variable.
predictions: An input column that contains the predicted class labels.

Arguments targets and predictions must be set to input columns of the same data type, one of the following: INTEGER, BOOLEAN, or CHAR/VARCHAR. Depending on their data type, these columns identify classes as follows:

INTEGER: Zero-based consecutive integers between 0 and (num-classes-1) inclusive, where num-classes is the number of classes. For example, given the following input column values— {0, 1, 2, 3, 4}—Vertica assumes five classes.

Note
If input column values are not consecutive, Vertica interpolates the missing values. Thus, given the following input values— {0, 1, 3, 5, 6,}— Vertica assumes seven classes.
BOOLEAN: Yes or No
CHAR/VARCHAR: Class names. If the input columns are of type CHAR/VARCHAR columns, you must also set parameter num_classes to the number of classes.
Note

Vertica computes the number of classes as the union of values in both input columns. For example, given the following sets of values in the targets and predictions input columns, Vertica counts four classes:
```
{'milk', 'soy milk', 'cream'}
{'soy milk', 'almond milk'}
```

Parameters

num_classes

An integer > 1, specifies the number of classes to pass to the function.

You must set this parameter if the specified input columns are of type CHAR/VARCHAR. Otherwise, the function processes this parameter according to the column data types:

INTEGER: By default set to 2, you must set this parameter correctly if the number of classes is any other value.
BOOLEAN: By default set to 2, cannot be set to any other value.

Examples

This example computes the confusion matrix for a logistic regression model that classifies cars in the mtcars data set as automatic or manual transmission. Observed values are in input column obs, while predicted values are in input column pred. Because this is a binary classification problem, all values are either 0 or 1.

In the table returned, all 19 cars with a value of 0 in column am are correctly predicted by PREDICT_LOGISTIC_REGRESSION as having a value of 0. Of the 13 cars with a value of 1 in column am, 12 are correctly predicted to have a value of 1, while 1 car is incorrectly classified as having a value of 0:

=> SELECT CONFUSION_MATRIX(obs::int, pred::int USING PARAMETERS num_classes=2) OVER()
    FROM (SELECT am AS obs, PREDICT_LOGISTIC_REG(mpg, cyl, disp,drat, wt, qsec, vs, gear, carb
             USING PARAMETERS model_name='myLogisticRegModel')AS PRED
             FROM mtcars) AS prediction_output;

actual_class | predicted_0 | predicted_1 |        comment
-------------+-------------+-------------+------------------------------------------
0            |          19 |           0 |
1            |           0 |          13 | Of 32 rows, 32 were used and 0 were ignored
(2 rows)

3.2 - CROSS_VALIDATE

Performs k-fold cross validation on a learning algorithm using an input relation, and grid search for hyper parameters.

Performs k-fold cross validation on a learning algorithm using an input relation, and grid search for hyper parameters. The output is an average performance indicator of the selected algorithm. This function supports SVM classification, naive bayes, and logistic regression.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

CROSS_VALIDATE ( 'algorithm', 'input-relation', 'response-column', 'predictor-columns'
        [ USING PARAMETERS
              [exclude_columns = 'excluded-columns']
           [, cv_model_name = 'model']
           [, cv_metrics = 'metrics']
           [, cv_fold_count = num-folds]
           [, cv_hyperparams = 'hyperparams']
           [, cv_prediction_cutoff = prediction-cutoff] ] )

Arguments

algorithm

Name of the algorithm training function, one of the following:

input-relation

The table or view that contains data used for training and testing. If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.

response-column

Name of the input column that contains the response.

predictor-columns

Parameters

exclude_columns

Comma-separated list of columns from predictor-columns to exclude from processing.

cv_model_name

The name of a model that lets you retrieve results of the cross validation process. If you omit this parameter, results are displayed but not saved. If you set this parameter to a model name, you can retrieve the results with summary functions GET_MODEL_ATTRIBUTE and GET_MODEL_SUMMARY

cv_metrics

The metrics used to assess the algorithm, specified either as a comma-separated list of metric names or in a JSON array. In both cases, you specify one or more of the following metric names:

accuracy (default)
error_rate
TP: True positive, the number of cases of class 1 predicted as class 1
FP: False positive, the number of cases of class 0 predicted as class 1
TN: True negative, the number of cases of class 0 predicted as class 0
FN: False negative, the number of cases of class 1 predicted as class 0
TPR or recall: True positive rate, the correct predictions among class 1
FPR: False positive rate, the wrong predictions among class 0
TNR: True negative rate, the correct predictions among class 0
FNR: False negative rate, the wrong predictions among class 1
PPV or precision: The positive predictive value, the correct predictions among cases predicted as class 1
NPV: Negative predictive value, the correct predictions among cases predicted as class 0
MSE: Mean squared error
MAE: Mean absolute error
rsquared: coefficient of determination
explained_variance

fscore

(1 + beta˄2) * precison * recall / (beta˄2 * precision + recall)

beta equals 1 by default

auc_roc: AUC of ROC using the specified number of bins, by default 100
auc_prc: AUC of PRC using the specified number of bins, by default 100
counts: Shortcut that resolves to four other metrics: TP, FP, TN, and FN
count: Valid only in JSON syntax, counts the number of cases labeled by one class (case-class-label) but predicted as another class (predicted-class-label):
```
cv_metrics='[{"count":[case-class-label, predicted-class-label]}]'
```

cv_fold_count

The number of folds to split the data.

Default: 5

cv_hyperparams

A JSON string that describes the combination of parameters for use in grid search of hyper parameters. The JSON string contains pairs of the hyper parameter name. The value of each hyper parameter can be specified as an array or sequence. For example:

{"param1":[value1,value2,...], "param2":{"first":first_value, "step":step_size, "count":number_of_values} }

Hyper parameter names and string values should be quoted using the JSON standard. These parameters are passed to the training function.

cv_prediction_cutoff

The cutoff threshold that is passed to the prediction stage of logistic regression, a FLOAT between 0 and 1, exclusive

Default: 0.5

Model attributes

call_string

The value of all input arguments that were specified at the time CROSS_VALIDATE was called.

run_average

The average across all folds of all metrics specified in parameter cv_metrics, if specified; otherwise, average accuracy.

fold_info

The number of rows in each fold:

fold_id: The index of the fold.
row_count: The number of rows held out for testing in the fold.

counters

All counters for the function, including:

accepted_row_count: The total number of rows in the input_relation, minus the number of rejected rows.
rejected_row_count: The number of rows of the input_relation that were skipped because they contained an invalid value.
feature_count: The number of features input to the machine learning model.

run_details

Information about each run, where a run means training a single model, and then testing that model on the one held-out fold:

fold_id: The index of the fold held out for testing.
iteration_count: The number of iterations used in model training on non-held-out folds.
accuracy: All metrics specified in parameter cv_metrics, or accuracy if cv_metrics is not provided.
error_rate: All metrics specified in parameter cv_metrics, or accuracy if the parameter is omitted.

Privileges

Non-superusers:

SELECT privileges on the input relation
CREATE and USAGE privileges on the default schema where machine learning algorithms generate models. If cv_model_name is provided, the cross validation results are saved as a model in the same schema.

Specifying metrics in JSON

Parameter cv_metrics can specify metrics as an array of JSON objects, where each object specifies a metric name . For example, the following expression sets cv_metrics to two metrics specified as JSON objects, accuracy and error_rate:

cv_metrics='["accuracy", "error_rate"]'

In the next example, cv_metrics is set to two metrics, accuracy and TPR (true positive rate). Here, the TPR metric is specified as a JSON object that takes an array of two class label arguments, 2 and 3:

cv_metrics='[ "accuracy", {"TPR":[2,3] } ]'

Metrics specified as JSON objects can accept parameters. In the following example, the fscore metric specifies parameter beta, which is set to 0.5:

cv_metrics='[ {"fscore":{"beta":0.5} } ]'

Parameter support can be especially useful for certain metrics. For example, metrics auc_roc and auc_prc build a curve, and then compute the area under that curve. For ROC, the curve is formed by plotting metrics TPR against FPR; for PRC, PPV (precision) against TPR (recall). The accuracy of such curves can be increased by setting parameter num_bins to a value greater than the default value of 100. For example, the following expression computes AUC for an ROC curve built with 1000 bins:

cv_metrics='[{"auc_roc":{"num_bins":1000}}]'

Using metrics with Multi-class classifier functions

All supported metrics are defined for binary classifier functions LOGISTIC_REG and SVM_CLASSIFIER. For multi-class classifier functions such as NAIVE_BAYES, these metrics can be calculated for each one-versus-the-rest binary classifier. Use arguments to request the metrics for each classifier. For example, if training data has integer class labels, you can set cv_metrics with the precision (PPV) metric as follows:

cv_metrics='[{"precision":[0,4]}]'

This setting specifies to return two columns with precision computed for two classifiers:

Column 1: classifies 0 versus not 0
Collumn 2: classifies 4 versus not 4

If you omit class label arguments, the class with index 1 is used. Instead of computing metrics for individual one-versus-the-rest classifiers, the average is computed in one of the following styles: macro, micro, or weighted (default). For example, the following cv_metrics setting returns the average weighted by class sizes:

cv_metrics='[{"precision":{"avg":"weighted"}}]'

AUC-type metrics can be similarly defined for multi-class classifiers. For example, the following cv_metrics setting computes the area under the ROC curve for each one-versus-the-rest classifier, and then returns the average weighted by class sizes.

cv_metrics='[{"auc_roc":{"avg":"weighted", "num_bins":1000}}]'

Examples

=> SELECT CROSS_VALIDATE('svm_classifier', 'mtcars', 'am', 'mpg'
      USING PARAMETERS cv_fold_count= 6,
                       cv_hyperparams='{"C":[1,5]}',
                       cv_model_name='cv_svm',
                       cv_metrics='accuracy, error_rate');
         CROSS_VALIDATE
----------------------------
 Finished

===========
run_average
===========
C  |accuracy      |error_rate
---+--------------+----------
1 | 0.75556       |  0.24444
5 | 0.78333       |  0.21667
(1 row)

3.3 - ERROR_RATE

Using an input table, returns a table that calculates the rate of incorrect classifications and displays them as FLOAT values.

Using an input table, returns a table that calculates the rate of incorrect classifications and displays them as FLOAT values. ERROR_RATE returns a table with the following dimensions:

Rows: Number of classes plus one row that contains the total error rate across classes
Columns: 2

Syntax

ERROR_RATE ( targets, predictions [ USING PARAMETERS num_classes = num-classes ] ) OVER()

Arguments

targets: An input column that contains the true values of the response variable.
predictions: An input column that contains the predicted class labels.

INTEGER: Zero-based consecutive integers between 0 and (num-classes-1) inclusive, where num-classes is the number of classes. For example, given the following input column values— {0, 1, 2, 3, 4}—Vertica assumes five classes.

Note
If input column values are not consecutive, Vertica interpolates the missing values. Thus, given the following input values— {0, 1, 3, 5, 6,}— Vertica assumes seven classes.
BOOLEAN: Yes or No
CHAR/VARCHAR: Class names. If the input columns are of type CHAR/VARCHAR columns, you must also set parameter num_classes to the number of classes.
Note

Vertica computes the number of classes as the union of values in both input columns. For example, given the following sets of values in the targets and predictions input columns, Vertica counts four classes:
```
{'milk', 'soy milk', 'cream'}
{'soy milk', 'almond milk'}
```

Parameters

num_classes

An integer > 1, specifies the number of classes to pass to the function.

You must set this parameter if the specified input columns are of type CHAR/VARCHAR. Otherwise, the function processes this parameter according to the column data types:

INTEGER: By default set to 2, you must set this parameter correctly if the number of classes is any other value.
BOOLEAN: By default set to 2, cannot be set to any other value.

Privileges

Non-superusers: model owner, or USAGE privileges on the model

Examples

This example shows how to execute the ERROR_RATE function on an input table named mtcars. The response variables appear in the column obs, while the prediction variables appear in the column pred. Because this example is a classification problem, all response variable values and prediction variable values are either 0 or 1, indicating binary classification.

In the table returned by the function, the first column displays the class id column. The second column displays the corresponding error rate for the class id. The third column indicates how many rows were successfully used by the function and whether any rows were ignored.

=> SELECT ERROR_RATE(obs::int, pred::int USING PARAMETERS num_classes=2) OVER()
    FROM (SELECT am AS obs, PREDICT_LOGISTIC_REG (mpg, cyl, disp, drat, wt, qsec, vs, gear, carb
                USING PARAMETERS model_name='myLogisticRegModel', type='response') AS pred
             FROM mtcars) AS prediction_output;
 class |     error_rate     |                   comment
-------+--------------------+---------------------------------------------
     0 |                  0 |
     1 | 0.0769230797886848 |
       |            0.03125 | Of 32 rows, 32 were used and 0 were ignored
(3 rows)

3.4 - LIFT_TABLE

Returns a table that compares the predictive quality of a machine learning model.

Returns a table that compares the predictive quality of a machine learning model. This function is also known as a lift chart.

Syntax

LIFT_TABLE ( targets, probabilities
        [ USING PARAMETERS [num_bins = num-bins] [, main_class = class-name ] ] )
OVER()

Arguments

targets

An input column that contains the true values of the response variable, one of the following data types: INTEGER, BOOLEAN, or CHAR/VARCHAR. Depending on the column data type, the function processes column data as follows:

INTEGER: Uses the input column as containing the true value of the response variable.
BOOLEAN: Resolves Yes to 1, 0 to No.
CHAR/VARCHAR: Resolves the value specified by parameter main_class to 1, all other values to 0.

Note

If the input column is of data type INTEGER or BOOLEAN, the function ignores parameter main_class.

probabilities

A FLOAT input column that contains the predicted probability of response being the main class, set to 1 if targets is of type INTEGER.

Parameters

num_bins

An integer value that determines the number of decision boundaries. Decision boundaries are set at equally spaced intervals between 0 and 1, inclusive. The function computes the table at each num-bin + 1 point.

Default: 100

main_class

Used only if targets is of type CHAR/VARCHAR, specifies the class to associate with the probabilities argument.

Examples

Execute LIFT_TABLE on an input table mtcars.

=> SELECT LIFT_TABLE(obs::int, prob::float USING PARAMETERS num_bins=2) OVER()
    FROM (SELECT am AS obs, PREDICT_LOGISTIC_REG(mpg, cyl, disp, drat, wt, qsec, vs, gear, carb
                                                    USING PARAMETERS model_name='myLogisticRegModel',
                                                    type='probability') AS prob
             FROM mtcars) AS prediction_output;
 decision_boundary | positive_prediction_ratio |       lift       |                   comment
-------------------+---------------------------+------------------+---------------------------------------------
                 1 |                         0 |              NaN |
               0.5 |                   0.40625 | 2.46153846153846 |
                 0 |                         1 |                1 | Of 32 rows, 32 were used and 0 were ignored
(3 rows)

The first column, decision_boundary, indicates the cut-off point for whether to classify a response as 0 or 1. For instance, for each row, if prob is greater than or equal to decision_boundary, the response is classified as 1. If prob is less than decision_boundary, the response is classified as 0.

The second column, positive_prediction_ratio, shows the percentage of samples in class 1 that the function classified correctly using the corresponding decision_boundary value.

For the third column, lift, the function divides the positive_prediction_ratio by the percentage of rows correctly or incorrectly classified as class 1.

3.5 - MSE

Returns a table that displays the mean squared error of the prediction and response columns in a machine learning model.

Syntax

MSE ( targets, predictions ) OVER()

Arguments

targets: The model response variable, of type FLOAT.
predictions: A FLOAT input column that contains predicted values for the response variable.

Examples

Execute the MSE function on input table faithful_testing. The response variables appear in the column obs, while the prediction variables appear in the column prediction.

=> SELECT MSE(obs, prediction) OVER()
   FROM (SELECT eruptions AS obs,
                PREDICT_LINEAR_REG (waiting USING PARAMETERS model_name='myLinearRegModel') AS prediction
         FROM faithful_testing) AS prediction_output;
        mse        |                   Comments
-------------------+-----------------------------------------------
 0.252925741352641 | Of 110 rows, 110 were used and 0 were ignored
(1 row)

3.6 - PRC

Returns a table that displays the points on a receiver precision recall (PR) curve.

Syntax

PRC ( targets, probabilities
       [ USING PARAMETERS
             [num_bins = num-bins]
             [, f1_score = return-score ]
             [, main_class = class-name ] )
OVER()

Arguments

targets

INTEGER: Uses the input column as containing the true value of the response variable.
BOOLEAN: Resolves Yes to 1, 0 to No.
CHAR/VARCHAR: Resolves the value specified by parameter main_class to 1, all other values to 0.

Note

If the input column is of data type INTEGER or BOOLEAN, the function ignores parameter main_class.

probabilities

A FLOAT input column that contains the predicted probability of response being the main class, set to 1 if targets is of type INTEGER.

Parameters

num_bins

Default: 100

f1_score

A Boolean that specifies whether to return a column that contains the f1 score—the harmonic average of the precision and recall measures, where an F1 score reaches its best value at 1 (perfect precision and recall) and worst at 0.

Default: false

main_class

Used only if targets is of type CHAR/VARCHAR, specifies the class to associate with the probabilities argument.

Examples

Execute the PRC function on an input table named mtcars. The response variables appear in the column obs, while the prediction variables appear in column pred.

=> SELECT PRC(obs::int, prob::float USING PARAMETERS num_bins=2, f1_score=true) OVER()
    FROM (SELECT am AS obs,
                    PREDICT_LOGISTIC_REG (mpg, cyl, disp, drat, wt, qsec, vs, gear, carb
                          USING PARAMETERS model_name='myLogisticRegModel',
                                           type='probability') AS prob
             FROM mtcars) AS prediction_output;
decision_boundary | recall | precision |     f1_score      |     comment
------------------+--------+-----------+-------------------+--------------------------------------------
0                 |      1 |   0.40625 | 0.577777777777778 |
0.5               |      1 |         1 |                 1 | Of 32 rows, 32 were used and 0 were ignored
(2 rows)

The first column, decision_boundary, indicates the cut-off point for whether to classify a response as 0 or 1. For example, in each row, if the probability is equal to or greater than decision_boundary, the response is classified as 1. If the probability is less than decision_boundary, the response is classified as 0.

3.7 - READ_TREE

Reads the contents of trees within the random forest or XGBoost model.

Syntax

READ_TREE ( USING PARAMETERS model_name = 'model-name' [, tree_id = tree-id] [, format = 'format'] )

Parameters

model_name

Identifies the model that is stored as a result of training, where model-name conforms to conventions described in Identifiers. It must also be unique among all names of sequences, tables, projections, views, and models within the same schema.

tree_id

The tree identifier, an integer between 0 and n-1, where n is the number of trees in the random forest or XGBoost model. If you omit this parameter, all trees are returned.

format

Output format of the returned tree, one of the following:

tabular: Returns a table with the twelve output columns.
graphviz: Returns DOT language source that can be passed to a graphviz tool and render a graphic visualization of the tree.

Privileges

Non-superusers: USAGE privileges on the model

Examples

Get tabular output from READ_TREE for a random forest model:

=> SELECT READ_TREE ( USING PARAMETERS model_name='myRFModel', tree_id=1 ,
format= 'tabular') LIMIT 2;
-[ RECORD 1 ]-------------+-------------------
tree_id                   | 1
node_id                   | 1
node_depth                | 0
is_leaf                   | f
is_categorical_split      | f
split_predictor           | petal_length
split_value               | 1.921875
weighted_information_gain | 0.111242236024845
left_child_id             | 2
right_child_id            | 3
prediction                |
probability/variance      |

-[ RECORD 2 ]-------------+-------------------
tree_id                   | 1
node_id                   | 2
node_depth                | 1
is_leaf                   | t
is_categorical_split      |
split_predictor           |
split_value               |
weighted_information_gain |
left_child_id             |
right_child_id            |
prediction                | setosa
probability/variance      | 1

Get graphviz-formatted output from READ_TREE:

=> SELECT READ_TREE ( USING PARAMETERS model_name='myRFModel', tree_id=1 ,
format= 'graphviz')LIMIT 1;

-[ RECORD 1 ]+-------------------------------------------------------------------
---------------------------------------------------------------------------------
tree_id      | 1
tree_digraph | digraph Tree{
1 [label="petal_length < 1.921875 ?", color="blue"];
1 -> 2 [label="yes", color="black"];
1 -> 3 [label="no", color="black"];
2 [label="prediction: setosa, probability: 1", color="red"];
3 [label="petal_length < 4.871875 ?", color="blue"];
3 -> 6 [label="yes", color="black"];
3 -> 7 [label="no", color="black"];
6 [label="prediction: versicolor, probability: 1", color="red"];
7 [label="prediction: virginica, probability: 1", color="red"];
}

This renders as follows:

3.8 - RF_PREDICTOR_IMPORTANCE

Measures the importance of the predictors in a random forest model using the Mean Decrease Impurity (MDI) approach.

Measures the importance of the predictors in a random forest model using the Mean Decrease Impurity (MDI) approach. The importance vector is normalized to sum to 1.

Syntax

RF_PREDICTOR_IMPORTANCE ( USING PARAMETERS model_name = 'model-name' [, tree_id = tree-id] )

Parameters

model_name: Identifies the model that is stored as a result of the training, where model-name must be of type rf_classifier or rf_regressor.
tree_id: Identifies the tree to process, an integer between 0 and n-1, where n is the number of trees in the forest. If you omit this parameter, the function uses all trees to measure importance values.

Privileges

Non-superusers: USAGE privileges on the model

Examples

This example shows how you can use the RF_PREDICTOR_IMPORTANCE function.

=> SELECT RF_PREDICTOR_IMPORTANCE ( USING PARAMETERS model_name = 'myRFModel');
 predictor_index | predictor_name | importance_value
-----------------+----------------+--------------------
               0 | sepal.length   | 0.106763318092655
               1 | sepal.width    | 0.0279536658041994
               2 | petal.length   | 0.499198722346586
               3 | petal.width    | 0.366084293756561
(4 rows)

3.9 - ROC

Returns a table that displays the points on a receiver operating characteristic curve.

Returns a table that displays the points on a receiver operating characteristic curve. The ROC function tells you the accuracy of a classification model as you raise the discrimination threshold for the model.

Syntax

ROC ( targets, probabilities
        [ USING PARAMETERS
              [num_bins = num-bins]
              [, AUC = output]
              [, main_class = class-name ] ) ] )
OVER()

Arguments

targets

INTEGER: Uses the input column as containing the true value of the response variable.
BOOLEAN: Resolves Yes to 1, 0 to No.
CHAR/VARCHAR: Resolves the value specified by parameter main_class to 1, all other values to 0.

Note

If the input column is of data type INTEGER or BOOLEAN, the function ignores parameter main_class.

probabilities

A FLOAT input column that contains the predicted probability of response being the main class, set to 1 if targets is of type INTEGER.

Parameters

num_bins

Default: 100

Greater values result in more precise approximations of the AUC.

AUC

A Boolean value that specifies whether to output the area under the curve (AUC) value.

Default: True

main_class

Used only if targets is of type CHAR/VARCHAR, specifies the class to associate with the probabilities argument.

Examples

Execute ROC on input table mtcars. Observed class labels are in column obs, predicted class labels are in column prob:

=> SELECT ROC(obs::int, prob::float USING PARAMETERS num_bins=5, AUC = True) OVER()
    FROM (SELECT am AS obs,
          PREDICT_LOGISTIC_REG (mpg, cyl, disp, drat, wt, qsec, vs, gear, carb
               USING PARAMETERS
                  model_name='myLogisticRegModel', type='probability') AS prob
   FROM mtcars) AS prediction_output;
 decision_boundary | false_positive_rate | true_positive_rate | AUC |comment
-------------------+---------------------+--------------------+-----+-----------------------------------
0                  |                   1 |                  1 |     |
0.5                |                   0 |                  1 |     |
1                  |                   0 |                  0 |   1 | Of 32 rows,32 were used and 0 were ignoreded
(3 rows)

The function returns a table with the following results:

decision_boundary indicates the cut-off point for whether to classify a response as 0 or 1. In each row, if prob is equal to or greater than decision_boundary, the response is classified as 1. If prob is less than decision_boundary, the response is classified as 0.
false_positive_rate shows the percentage of false positives (when 0 is classified as 1) in the corresponding decision_boundary.
true_positive_rate shows the percentage of rows that were classified as 1 and also belong to class 1.

3.10 - RSQUARED

Returns a table with the R-squared value of the predictions in a regression model.

Syntax

RSQUARED ( targets, predictions ) OVER()

Important

The OVER() clause must be empty.

Arguments

targets: A FLOAT response variable for the model.
predictions: A FLOAT input column that contains the predicted values for the response variable.

Examples

This example shows how to execute the RSQUARED function on an input table named faithful_testing. The observed values of the response variable appear in the column, obs, while the predicted values of the response variable appear in the column, pred.

=> SELECT RSQUARED(obs, prediction) OVER()
     FROM (SELECT eruptions AS obs,
                  PREDICT_LINEAR_REG (waiting
                                       USING PARAMETERS model_name='myLinearRegModel') AS prediction
           FROM faithful_testing) AS prediction_output;
        rsq        |                    comment
-------------------+-----------------------------------------------
 0.801392981147911 | Of 110 rows, 110 were used and 0 were ignored
(1 row)

3.11 - XGB_PREDICTOR_IMPORTANCE

Measures the importance of the predictors in an XGBoost model.

Measures the importance of the predictors in an XGBoost model. The function outputs three measures of importance for each predictor:

frequency: relative number of times the model uses a predictor to split the data.
total_gain: relative contribution of a predictor to the model based on the total information gain across a predictor's splits. A higher value means more predictive importance.
avg_gain: relative contribution of a predictor to the model based on the average information gain across a predictor's splits.

The sum of each importance measure is normalized to one across all predictors.

Syntax

XGB_PREDICTOR_IMPORTANCE ( USING PARAMETERS param=value[,...] )

Parameters

model_name: Name of the model, which must be of type xgb_classifier or xgb_regressor.
tree_id: Integer in the range [0, n-1], where n is the number of trees in model_name, that specifies the tree to process. If you omit this parameter, the function uses all trees in the model to measure predictor importance values.

Privileges

Non-superusers: USAGE privileges on the model

Examples

The following example measures the importance of the predictors in the model 'xgb_iris', an XGBoost classifier model, across all trees:

=> SELECT XGB_PREDICTOR_IMPORTANCE( USING PARAMETERS model_name = 'xgb_iris' );
 predictor_index | predictor_name |     frequency     |     total_gain     |      avg_gain
-----------------+----------------+-------------------+--------------------+--------------------
               0 | sepal_length   |  0.15384615957737 |    0.0183021749937 | 0.0370849960701401
               1 | sepal_width    | 0.215384617447853 | 0.0154729501420881 | 0.0223944615251752
               2 | petal_length   | 0.369230777025223 |  0.607349886817728 |  0.512770753876444
               3 | petal_width    | 0.261538475751877 |  0.358874988046484 |  0.427749788528241
(4 rows)

To sort the predictors by importance values, you can use a nested query with an ORDER BY clause. The following sorts the model predictors by descending avg_gain:

=> SELECT * FROM (SELECT XGB_PREDICTOR_IMPORTANCE( USING PARAMETERS model_name = 'xgb_iris' )) AS importances ORDER BY avg_gain DESC;
 predictor_index | predictor_name |     frequency     |     total_gain     |      avg_gain
-----------------+----------------+-------------------+--------------------+--------------------
               2 | petal_length   | 0.369230777025223 |  0.607349886817728 |  0.512770753876444
               3 | petal_width    | 0.261538475751877 |  0.358874988046484 |  0.427749788528241
               0 | sepal_length   |  0.15384615957737 |    0.0183021749937 | 0.0370849960701401
               1 | sepal_width    | 0.215384617447853 | 0.0154729501420881 | 0.0223944615251752
(4 rows)

4 - Model management

Vertica provides several functions for managing models.

4.1 - CHANGE_MODEL_STATUS

Changes the status of a registered model.

Changes the status of a registered model. Only dbadmin and users with the MLSUPERVISOR role can call this function.

The following diagram depicts the valid status transitions:

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Stable

Syntax

CHANGE_MODEL_STATUS( 'registered_name', registered_version, 'new_status' )

Arguments

registered_name

Identifies the abstract name to which the model is registered. This registered_name can represent a group of models for a higher-level application, where each model in the group has a unique version number.

registered_version

Unique version number of the model under the specified registered_name.

If there is no registered model with the given registered_name and registered_version, the function errors.

new_status

New status of the registered model. Must be one of the following strings and adhere to the valid status transitions depicted in the above diagram:

under_review: Status assigned to newly registered models.
staging: Model is targeted for A/B testing against the model currently in production.
production: Model is in production for its specified application. Only one model can be in production for a given registered_name at one time.
archived: Status of models that were previously in production. Archived models can be returned to production at any time.
declined: Model is no longer in consideration for production.
unregistered: Model is removed from the versioning environment. The model does not appear in the REGISTERED_MODELS system table.

If you change the status of a model to 'production' and there is already a model in production under the given registered_name, the status of the model in production is set to 'archived' and the status of the new model is set to 'production'.

Privileges

One of the following:

Superuser
MLSUPERVISOR

Examples

In the following example, the linear_reg_spark1 model, which is uniquely identified by the registered_name 'linear_reg_app' and the registered_version of two, is set to 'production' status:

=> SELECT * FROM REGISTERED_MODELS;
  registered_name | registered_version |    status    |        registered_time        |      model_id     | schema_name |    model_name     |      model_type       |    category
------------------+--------------------+--------------+-------------------------------+-------------------+-------------+-------------------+-----------------------+----------------
 linear_reg_app   |                  2 | STAGING      | 2023-01-29 05:49:00.082166-04 | 45035996273714020 | public      | linear_reg_spark1 | PMML_REGRESSION_MODEL | PMML
 linear_reg_app   |                  1 | PRODUCTION   | 2023-01-24 09:19:04.553102-05 | 45035996273850350 | public      | native_linear_reg | LINEAR_REGRESSION     | VERTICA_MODELS
 logistic_reg_app |                  1 | DECLINED     | 2023-01-11 02:47:25.990626-02 | 45035996273853740 | public      | log_reg_bfgs      | LOGISTIC_REGRESSION   | VERTICA_MODELS
(3 rows)

=> SELECT CHANGE_MODEL_STATUS('linear_reg_app', 2, 'production');
                          CHANGE_MODEL_STATUS
-----------------------------------------------------------------------------
The status of model [linear_reg_app] - version [2] is changed to [production]
(1 row)

You can query the REGISTERED_MODELS system table to confirm that the linear_reg_spark1 model is now in 'production' and the native_linear_reg model, which was currently in 'production', is moved to 'archived':

=> SELECT * FROM REGISTERED_MODELS;
  registered_name | registered_version |    status    |        registered_time        |      model_id     | schema_name |    model_name     |      model_type       |    category
------------------+--------------------+--------------+-------------------------------+-------------------+-------------+-------------------+-----------------------+----------------
 linear_reg_app   |                  2 | PRODUCTION   | 2023-01-29 05:49:00.082166-04 | 45035996273714020 | public      | linear_reg_spark1 | PMML_REGRESSION_MODEL | PMML
 linear_reg_app   |                  1 | ARCHIVED     | 2023-01-24 09:19:04.553102-05 | 45035996273850350 | public      | native_linear_reg | LINEAR_REGRESSION     | VERTICA_MODELS
 logistic_reg_app |                  1 | DECLINED     | 2023-01-11 02:47:25.990626-02 | 45035996273853740 | public      | log_reg_bfgs      | LOGISTIC_REGRESSION   | VERTICA_MODELS
(2 rows)

If you change a model's status to 'unregistered', the model is removed from the model versioning environment and no longer appears in the REGISTERED_MODELS system table:

=> SELECT CHANGE_MODEL_STATUS('logistic_reg_app', 1, 'unregistered');
                            CHANGE_MODEL_STATUS
----------------------------------------------------------------------------------
The status of model [logistic_reg_app] - version [1] is changed to [unregistered]
(1 row)

=> SELECT * FROM REGISTERED_MODELS;
  registered_name | registered_version |    status    |        registered_time        |      model_id     | schema_name |    model_name     |      model_type       |    category
------------------+--------------------+--------------+-------------------------------+-------------------+-------------+-------------------+-----------------------+----------------
 linear_reg_app   |                  2 | STAGING      | 2023-01-29 05:49:00.082166-04 | 45035996273714020 | public      | linear_reg_spark1 | PMML_REGRESSION_MODEL | PMML
 linear_reg_app   |                  1 | PRODUCTION   | 2023-01-24 09:19:04.553102-05 | 45035996273850350 | public      | native_linear_reg | LINEAR_REGRESSION     | VERTICA_MODELS
(2 rows)

4.2 - EXPORT_MODELS

Exports machine learning models.

Exports machine learning models. Vertica supports three model formats:

Native Vertica (VERTICA_MODELS)
PMML
TensorFlow

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

EXPORT_MODELS ( 'output-dir', 'export-target' [ USING PARAMETERS category = 'model-category' ] )

Arguments

output-dir

Absolute path of an output directory to store the exported models.

export-target

Specifies which models to export as follows:

 [schema.]{ model-name | * }

where schema optionally specifies to export models from the specified schema. If omitted, EXPORT_MODELS uses the default schema. Supply * (asterisk) to export all models from the schema.

Parameters

category

The category of models to export, one of the following:

VERTICA_MODELS
PMML
TENSORFLOW

EXPORT_MODELS exports models of the specified category according to the scope of the export operation—that is, whether it applies to a single model, or to all models within a schema. See Export Scope and Category Processing below.

Exported Files below describes the files that EXPORT_MODELS exports for each category.

If you omit this parameter, EXPORT_MODELS exports the model, or models in the specified schema, according to their model type.

Privileges

Superuser

Export scope and category processing

EXPORT_MODELS executes according to the following parameter settings:

Scope of the export operation: single model, or all models within a given schema
Category specified or omitted

The following table shows how these two parameters control the export process:

Export scope	If category specified...	If category omitted...
Single model	Convert the model to the specified category, provided the model and category are compatible; otherwise, return with a mismatch error.	Export the model according to model type.
All models in schema	Export only models that are compatible with the specified category and issue mismatch warnings on all other models in the schema.	Export all models in the schema according to model type.

Exported files

EXPORT_MODELS exports the following files for each model category:

Model category	Exported files
`VERTICA_MODELS`	Multiple binary files (exact number dependent on model type) `metadata.json`: Metadata file with model information —model name, category, type, Vertica version on export. `crc.json`: Used on import to validate other files of this model.
`PMML`	XML file with the same name as the model and complying with PMML standard. `metadata.json`: Metadata file with model information —model name, category, type, Vertica version on export. `crc.json`: Used on import to validate other files of this model.
`TENSORFLOW`	`model-name``.pb`: Contains the TensorFlow model, saved in 'frozen graph' format. `metadata.json`: Metadata file with model information —model name, category, type, Vertica version on export. `tf_model_desc.json`: Summary model description. `model.json`: Verbose model description. `crc.json`: Used on import to validate other files of this model.

Categories and compatible models

If EXPORT_MODELS specifies a single model and also sets the category parameter, the function succeeds if the model type and category are compatible; otherwise, it returns with an error:

Model type	Compatible categories
PMML	`PMML`
TensorFlow	`TENSORFLOW`
VERTICA_MODELS	`PMML VERTICA_MODELS`

If EXPORT_MODELS specifies to export all models from a schema and sets a category, it issues a warning message on each model that is incompatible with that category. The function then continues to process remaining models in that schema.

EXPORT_MODELS logs all errors and warnings in output-dir/export_log.json.

Examples

Export models without changing their category:

Export model myschema.mykmeansmodel without changing its category:

=> SELECT EXPORT_MODELS ('/home/dbadmin', 'myschema.mykmeansmodel');
EXPORT_MODELS
----------------
Success
(1 row)

Export all models in schema myschema without changing their categories:

=> SELECT EXPORT_MODELS ('/home/dbadmin', 'myschema.*');
EXPORT_MODELS
----------------
Success
(1 row)

Export models that are compatible with the specified category:

Note

When you import a model of category VERTICA_MODELS trained in a different version of Vertica, Vertica automatically upgrades the model version to match that of the database. If this fails, you must run UPGRADE_MODEL.

If both methods fail, the model cannot be used for in-database scoring and cannot be exported as a PMML model.

The category is set to PMML. Models of type PMML and VERTICA_MODELS are compatible with the PMML category, so the export operation succeeds if my_keans is of either type:
```
=> SELECT EXPORT_MODELS ('/tmp/', 'my_kmeans' USING PARAMETERS category='PMML');
```
The category is set to VERTICA_MODELS. Only models of type VERTICA_MODELS are compatible with the VERTICA_MODELS category, so the export operation succeeds only if my_keans is of that type:
```
=> SELECT EXPORT_MODELS ('/tmp/', 'public.my_kmeans' USING PARAMETERS category='VERTICA_MODELS');
```
The category is set to TENSORFLOW. Only models of type TensorFlow are compatible with the TENSORFLOW category, so the model tf_mnist_keras must be of type TensorFlow:
```
=> SELECT EXPORT_MODELS ('/tmp/', 'tf_mnist_keras', USING PARAMETERS category='TENSORFLOW');
export_models
---------------
Success
(1 row)
```

After exporting the TensorFlow model tf_mnist_keras, list the exported files:

$ ls tf_mnist_keras/
crc.json  metadata.json  mnist_keras.pb  model.json  tf_model_desc.json

4.3 - GET_MODEL_ATTRIBUTE

Extracts either a specific attribute from a model or all attributes from a model.

Extracts either a specific attribute from a model or all attributes from a model. Use this function to view a list of attributes and row counts or view detailed information about a single attribute. The output of GET_MODEL_ATTRIBUTE is a table format where users can select particular columns or rows.

Syntax

GET_MODEL_ATTRIBUTE ( USING PARAMETERS model_name = 'model-name' [, attr_name = 'attribute' ] )

Parameters

model_name: Name of the model (case-insensitive).
attr_name: Name of the model attribute to extract. If omitted, the function shows all available attributes. Attribute names are case-sensitive.

Privileges

Non-superusers: model owner, or USAGE privileges on the model

Examples

This example returns a summary of all model attributes.

=> SELECT GET_MODEL_ATTRIBUTE ( USING PARAMETERS model_name='myLinearRegModel');
attr_name          |                attr_fields                        | #_of_rows
-------------------+---------------------------------------------------+-----------
details            | predictor, coefficient, std_err, t_value, p_value |         2
regularization     | type, lambda                                      |         1
iteration_count    | iteration_count                                   |         1
rejected_row_count | rejected_row_count                                |         1
accepted_row_count | accepted_row_count                                |         1
call_string        | call_string                                       |         1
(6 rows)

This example extracts the details attribute from the myLinearRegModel model.

=> SELECT GET_MODEL_ATTRIBUTE ( USING PARAMETERS model_name='myLinearRegModel', attr_name='details');
coeffNames |       coeff        |       stdErr        |      zValue       |        pValue
-----------+--------------------+---------------------+-------------------+-----------------------
Intercept  |  -1.87401598641074 |   0.160143331525544 | -11.7021169008952 |   7.3592939615234e-26
waiting    | 0.0756279479518627 | 0.00221854185633525 |  34.0890336307608 | 8.13028381124448e-100
(2 rows)

4.4 - GET_MODEL_SUMMARY

Returns summary information of a model.

Syntax

GET_MODEL_SUMMARY ( USING PARAMETERS model_name = 'model-name' )

Parameters

model_name: Name of the model (case-insensitive).

Privileges

Non-superusers: model owner, or USAGE privileges on the model

Examples

This example shows how you can view the summary of a linear regression model.

=> SELECT GET_MODEL_SUMMARY( USING PARAMETERS model_name='myLinearRegModel');

--------------------------------------------------------------------------------
=======
details
=======
predictor|coefficient|std_err |t_value |p_value
---------+-----------+--------+--------+--------
Intercept| -2.06795  | 0.21063|-9.81782| 0.00000
waiting  |  0.07876  | 0.00292|26.96925| 0.00000

==============
regularization
==============
type| lambda
----+--------
none| 1.00000

===========
call_string
===========
linear_reg('public.linear_reg_faithful', 'faithful_training', '"eruptions"', 'waiting'
USING PARAMETERS optimizer='bfgs', epsilon=1e-06, max_iterations=100,
regularization='none', lambda=1)

===============
Additional Info
===============
Name              |Value
------------------+-----
iteration_count   |  3
rejected_row_count|  0
accepted_row_count| 162
(1 row)

4.5 - IMPORT_MODELS

Imports models into Vertica, either Vertica models that were exported with EXPORT_MODELS, or models in Predictive Model Markup Language (PMML) or TensorFlow format.

Imports models into Vertica, either Vertica models that were exported with EXPORT_MODELS, or models in Predictive Model Markup Language (PMML) or TensorFlow format. You can use this function to move models between Vertica clusters, or to import PMML and TensorFlow models trained elsewhere.

Other Vertica model management operations such as GET_MODEL_SUMMARY and GET_MODEL_ATTRIBUTE support imported models.

Caution

Changing the exported model files causes the import functionality to fail on attempted re-import.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

IMPORT_MODELS ( 'source'
           [ USING PARAMETERS [ new_schema = 'schema-name' ] [, category = 'model-category' ] ] )

Arguments

source

The absolute path of the location from which to import models, one of the following:

The directory of a single model:
```
path/model-directory
```
The parent directory of multiple model directories:
```
parent-dir-path/*
```

Parameters

new_schema

An existing schema where the machine learning models are imported. If omitted, models are imported to the default schema.

IMPORT_MODELS extracts the name of the imported model from its metadata.json file, if it exists. Otherwise, the function uses the name of the model directory.

category

Specifies the category of the model to import, one of the following:

VERTICA_MODELS
PMML
TENSORFLOW

This parameter is required if the model directory has no metadata.json file. IMPORT_MODELS returns with an error if one of the following cases is true:

No category is specified and the model directory has no metadata.json.
The specified category does not match the model type.

Note

If the category is TENSORFLOW, IMPORT_MODELS only imports the following files from the model directory:

model-name.pb
model-name.json
model-name.pbtxt (optional)

Privileges

Superuser

Requirements and restrictions

The following requirements and restrictions apply:

If you export a model, then import it again, the export and import model directory names must match. If naming conflicts occur, import the model to a different schema by using the new_schema parameter, and then rename the model.
The machine learning configuration parameter MaxModelSizeKB sets the maximum size of a model that can be imported into Vertica.
Some PMML features and attributes are not currently supported. See PMML features and attributes for details.
If you import a PMML model with both metadata.json and crc.json files, the CRC file must contain the metadata file's CRC value. Otherwise, the import operation returns with an error.

Examples

Import models into the specified schema:

In both examples no model category is specified, so IMPORT_MODEL uses the model's metadata.json file to determine its category:

Import a single model mykmeansmodel into the newschema schema:

=> SELECT IMPORT_MODELS ('/home/dbadmin/myschema/mykmeansmodel' USING PARAMETERS new_schema='newschema')
IMPORT_MODELS
----------------
Success
(1 row)

Import all models in the myschema directory into the newschema schema:

=> SELECT IMPORT_MODELS ('/home/dbadmin/myschema/*' USING PARAMETERS new_schema='newschema')
IMPORT_MODELS
----------------
Success
(1 row)

Specify the category of models to import:

In the first two examples, IMPORT_MODELS returns with success only if the specified model and category match; otherwise, it returns an error:

Import kmeans_pmml as a PMML model:

SELECT IMPORT_MODELS ('/root/user/kmeans_pmml' USING PARAMETERS category='PMML')
 import_models
---------------
 Success
(1 row)

Import tf_mnist_estimator as a TensorFlow model:

=> SELECT IMPORT_MODELS ( '/path/tf_models/tf_mnist_estimator' USING PARAMETERS category='TENSORFLOW');
 import_models
---------------
 Success
(1 row)

Import all TensorFlow models from the specified directory:

=> SELECT IMPORT_MODELS ( '/path/tf_models/*' USING PARAMETERS category='TENSORFLOW');
 import_models
---------------
 Success
(1 row)

4.6 - REGISTER_MODEL

Registers a trained model and adds it to Model Versioning environment with a status of 'under_review'.

Registers a trained model and adds it to Model versioning environment with a status of 'under_review'. The model must be registered by the owner of the model, dbadmin, or MLSUPERVISOR.

After a model is registered, the model owner is automatically changed to Superuser and the previous owner is given USAGE privileges. Users with the MLSUPERVISOR role or dbamin can call the CHANGE_MODEL_STATUS function to alter the status of registered models.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Stable

Syntax

REGISTER_MODEL( 'model_name', 'registered_name' )

Arguments

model_name: Identifies the model to register. If the model has already been registered, the function throws an error.
registered_name: Identifies an abstract name to which the model is registered. This registered_name can represent a group of models for a higher-level application, where each model in the group has a unique version number.
If a model is the first to be registered to a given registered_name, the model is assigned a registered_version of one. Otherwise, newly registered models are assigned an incremented registered_version of n + 1, where n is the number of models already registered to the given registered_name. Each registered model can be uniquely identified by the combination of registered_name and registered_version.

Privileges

Non-superusers: model owner

Examples

In the following example, the model log_reg_bfgs is registered to the logistic_reg_app application:

=> SELECT REGISTER_MODEL('log_reg_bfgs', 'logistic_reg_app');
                          REGISTER_MODEL
----------------------------------------------------------------------
Model [log_reg_bfgs] is registered as [logistic_reg_app], version [1]
(1 row)

You can query the REGISTERED_MODELS system table to view details about the newly registered model:

=> SELECT * FROM REGISTERED_MODELS;
  registered_name | registered_version |    status    |        registered_time        |      model_id     | schema_name |    model_name     |      model_type       |    category
------------------+--------------------+--------------+-------------------------------+-------------------+-------------+-------------------+-----------------------+----------------
 logistic_reg_app |                  1 | UNDER_REVIEW | 2023-01-22 09:49:25.990626-02 | 45035996273853740 | public      | log_reg_bfgs      | LOGISTIC_REGRESSION   | VERTICA_MODELS
(1 row)

4.7 - UPGRADE_MODEL

Upgrades a model from a previous Vertica version.

Upgrades a model from a previous Vertica version. Vertica automatically runs this function during a database upgrade and if you run the IMPORT_MODELS function. Manually call this function to upgrade models after a backup or restore.

If UPGRADE_MODEL fails to upgrade the model and the model is of category VERTICA_MODELS, it cannot be used for in-database scoring and cannot be exported as a PMML model.

This is a meta-function. You must call meta-functions in a top-level SELECT statement.

Behavior type

Volatile

Syntax

UPGRADE_MODEL ( [ USING PARAMETERS [model_name = 'model-name'] ] )

Parameters

model_name: Name of the model to upgrade. If you omit this parameter, Vertica upgrades all models on which you have privileges.

Privileges

Non-superuser: Upgrades only models that the user owns.

Examples

Upgrade model myLogisticRegModel:

=> SELECT UPGRADE_MODEL( USING PARAMETERS model_name = 'myLogisticRegModel');
        UPGRADE_MODEL
----------------------------
 1 model(s) upgrade

(1 row)

Upgrade all models that the user owns:

=> SELECT UPGRADE_MODEL();
        UPGRADE_MODEL
----------------------------
 20 model(s) upgrade

(1 row)

5 - Transformation functions

The machine learning API includes a set of UDx functions that transform the columns of each input row to one or more corresponding output columns.

The machine learning API includes a set of UDx functions that transform the columns of each input row to one or more corresponding output columns. These transformations follow rules that are defined in models that were created earlier. For example, APPLY_SVD uses an SVD model to transform input data.

Unless otherwise indicated, these functions require the following privileges for non-superusers:

USAGE privileges on the model
SELECT privileges on the input relation

In general, given an invalid input row, the return value for these functions is NULL.

5.1 - APPLY_BISECTING_KMEANS

Applies a trained bisecting k-means model to an input relation, and assigns each new data point to the closest matching cluster in the trained model.

Note

If the input relation is defined in Hive, use SYNC_WITH_HCATALOG_SCHEMA to sync the hcatalog schema, and then run the machine learning function.

Syntax

SELECT APPLY_BISECTING_KMEANS( 'input-columns'
        USING PARAMETERS model_name = 'model-name'
            [, num_clusters = 'num-clusters']
            [, match_by_pos = match-by-position] ] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. Input columns must be of data type numeric.

Parameters

model_name

Name of the model (case-insensitive).

num_clusters

Integer between 1 and k inclusive, where k is the number of centers in the model, specifies the number of clusters to use for prediction.

Default: Value that the model specifies for k

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Privileges

Non-superusers: model owner, or USAGE privileges on the model

5.2 - APPLY_IFOREST

Applies an isolation forest (iForest) model to an input relation.

Applies an isolation forest (iForest) model to an input relation. For each input row, the function returns an output row with two fields:

anomaly_score: A float value that represents the average path length across all trees in the model normalized by the training sample size.
is_anomaly: A Boolean value that indicates whether the input row is an anomaly. This value is true when anomaly_score is equal to or larger than a given threshold; otherwise, it's false.

Syntax

APPLY_IFOREST( input-columns USING PARAMETERS param=value[,...] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. Column types must match the types of the predictors in model_name.

Parameters

model_name

Name of the model (case-insensitive).

threshold

Optional. Float in the range (0.0, 1.0), specifies the threshold that determines if a data point is an anomaly. If the anomaly_score for a data point is equal to or larger than the value of threshold, the data point is marked as an outlier.

Alternatively, you can specify a contamination value that sets a threshold where the percentage of training data points labeled as outliers is approximately equal to the value of contamination. You cannot set both contamination and threshold in the same function call.

Default: 0.7

match_by_pos

Optional. Boolean value that specifies how input columns are matched to model columns:

false: Match by name.
true: Match by the position of columns in the input columns list.

Default: false

contamination

Optional. Float in the range (0.0, 1.0), the approximate ratio of data points in the training data that are labeled as outliers. The function calculates a threshold based on this contamination value. If you do not set this parameter, the function marks outliers using the specified or default threshold value.

You cannot set both contamination and threshold in the same function call.

Privileges

Non-superusers:

USAGE privileges on the model
SELECT privileges on the input relation

Examples

The following example demonstrates how different threshold values can affect outlier detection on an input relation:

=> SELECT * FROM (SELECT first_name, last_name, APPLY_IFOREST(team, hr, hits, avg, salary USING PARAMETERS model_name='baseball_anomalies',
   threshold=0.75) AS predictions FROM baseball) AS outliers WHERE predictions.is_anomaly IS true;
 first_name | last_name |                      predictions
------------+-----------+-------------------------------------------------------
 Jacqueline | Richards  | {"anomaly_score":0.777757463074347,"is_anomaly":true}
(1 row)

=> SELECT * FROM (SELECT first_name, last_name, APPLY_IFOREST(team, hr, hits, avg, salary USING PARAMETERS model_name='baseball_anomalies',
   threshold=0.55) AS predictions FROM baseball) AS outliers WHERE predictions.is_anomaly IS true;
 first_name | last_name |                      predictions
------------+-----------+--------------------------------------------------------
 Jacqueline | Richards  | {"anomaly_score":0.777757463074347,"is_anomaly":true}
 Debra      | Hall      | {"anomaly_score":0.5714649698133808,"is_anomaly":true}
 Gerald     | Fuller    | {"anomaly_score":0.5980549926114661,"is_anomaly":true}
(3 rows)

You can also use different contamination values to alter the outlier threshold:

=> SELECT * FROM (SELECT first_name, last_name, APPLY_IFOREST(team, hr, hits, avg, salary USING PARAMETERS model_name='baseball_anomalies',
   contamination = 0.1) AS predictions FROM baseball) AS outliers WHERE predictions.is_anomaly IS true;
 first_name | last_name |                      predictions
------------+-----------+--------------------------------------------------------
 Marie      | Fields    | {"anomaly_score":0.5307715717521868,"is_anomaly":true}
 Jacqueline | Richards  | {"anomaly_score":0.777757463074347,"is_anomaly":true}
 Debra      | Hall      | {"anomaly_score":0.5714649698133808,"is_anomaly":true}
 Gerald     | Fuller    | {"anomaly_score":0.5980549926114661,"is_anomaly":true}
(4 rows)

=> SELECT * FROM (SELECT first_name, last_name, APPLY_IFOREST(team, hr, hits, avg, salary USING PARAMETERS model_name='baseball_anomalies',
   contamination = 0.01) AS predictions FROM baseball) AS outliers WHERE predictions.is_anomaly IS true;
 first_name | last_name |                      predictions
------------+-----------+--------------------------------------------------------
 Jacqueline | Richards  | {"anomaly_score":0.777757463074347,"is_anomaly":true}
 Debra      | Hall      | {"anomaly_score":0.5714649698133808,"is_anomaly":true}
 Gerald     | Fuller    | {"anomaly_score":0.5980549926114661,"is_anomaly":true}
(3 rows)

5.3 - APPLY_INVERSE_PCA

Inverts the APPLY_PCA-generated transform back to the original coordinate system.

Syntax

APPLY_INVERSE_PCA ( input-columns
        USING PARAMETERS model_name = 'model-name'
            [, exclude_columns = 'excluded-columns']
            [, key_columns = 'key-columns'] )

Arguments

input-columns

Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. The following requirements apply:

All columns must be a numeric data type.
Enclose the column name in double quotes if it contains special characters.

Parameters

model_name: Name of the model (case-insensitive).
exclude_columns: Comma-separated list of column names from input-columns to exclude from processing.
key_columns: Comma-separated list of column names from input-columns that identify its data rows. These columns are included in the output table.

Examples

The following example shows how to use the APPLY_INVERSE_PCA function. It shows the output for the first record.


=> SELECT PCA ('pcamodel', 'world','country,HDI,em1970,em1971,em1972,em1973,em1974,em1975,em1976,em1977,
em1978,em1979,em1980,em1981,em1982,em1983,em1984 ,em1985,em1986,em1987,em1988,em1989,em1990,em1991,em1992,
em1993,em1994,em1995,em1996,em1997,em1998,em1999,em2000,em2001,em2002,em2003,em2004,em2005,em2006,em2007,
em2008,em2009,em2010,gdp1970,gdp1971,gdp1972,gdp1973,gdp1974,gdp1975,gdp1976,gdp1977,gdp1978,gdp1979,gdp1980,
gdp1981,gdp1982,gdp1983,gdp1984,gdp1985,gdp1986,gdp1987,gdp1988,gdp1989,gdp1990,gdp1991,gdp1992,gdp1993,
gdp1994,gdp1995,gdp1996,gdp1997,gdp1998,gdp1999,gdp2000,gdp2001,gdp2002,gdp2003,gdp2004,gdp2005,gdp2006,
gdp2007,gdp2008,gdp2009,gdp2010' USING PARAMETERS exclude_columns='HDI,country');
PCA
---------------------------------------------------------------
Finished in 1 iterations.
Accepted Rows: 96  Rejected Rows: 0
(1 row)
=> CREATE TABLE worldPCA AS SELECT
APPLY_PCA (HDI,country,em1970,em1971,em1972,em1973,em1974,em1975,em1976,em1977,em1978,em1979,
em1980,em1981,em1982,em1983,em1984 ,em1985,em1986,em1987,em1988,em1989,em1990,em1991,em1992,em1993,em1994,
em1995,em1996,em1997,em1998,em1999,em2000,em2001,em2002,em2003,em2004,em2005,em2006,em2007,em2008,em2009,
em2010,gdp1970,gdp1971,gdp1972,gdp1973,gdp1974,gdp1975,gdp1976,gdp1977,gdp1978,gdp1979,gdp1980,gdp1981,gdp1982,
gdp1983,gdp1984,gdp1985,gdp1986,gdp1987,gdp1988,gdp1989,gdp1990,gdp1991,gdp1992,gdp1993,gdp1994,gdp1995,
gdp1996,gdp1997,gdp1998,gdp1999,gdp2000,gdp2001,gdp2002,gdp2003,gdp2004,gdp2005,gdp2006,gdp2007,gdp2008,
gdp2009,gdp2010 USING PARAMETERS model_name='pcamodel', exclude_columns='HDI, country', key_columns='HDI,
country',cutoff=.3)OVER () FROM world;
CREATE TABLE

=> SELECT * FROM worldPCA;
HDI   |       country       |       col1
------+---------------------+-------------------
0.886 | Belgium             |  79002.2946705704
0.699 | Belize              | -25631.6670012556
0.427 | Benin               | -40373.4104598122
0.805 | Chile               | -16805.7940082156
0.687 | China               | -37279.2893141103
0.744 | Costa Rica          | -19505.5631231635
0.4   | Cote d'Ivoire       | -38058.2060339272
0.776 | Cuba                | -23724.5779612041
0.895 | Denmark             |  117325.594028813
0.644 | Egypt               | -34609.9941604549
...
(96 rows)

=> SELECT APPLY_INVERSE_PCA (HDI, country, col1
    USING PARAMETERS model_name = 'pcamodel', exclude_columns='HDI,country',
    key_columns = 'HDI, country') OVER () FROM worldPCA;
HDI  |       country       |      em1970       |      em1971       |      em1972      |      em1973      |
      em1974      |      em1975       |      em1976|      em1977      |      em1978       |      em1979
   |      em1980       |      em1981      |      em1982       |      em1983       |      em1984       |em1985
|      em1986       |      em1987       |      em1988       |      em1989      |      em1990      |      em1991
|      em1992       |      em1993|      em1994      |      em1995       |      em1996       |      em1997
    |      em1998       |      em1999       |      em2000       |      em2001       |em2002       |
em2003      |      em2004       |      em2005      |      em2006       |      em2007       |      em2008
|      em2009      |      em2010       |     gdp1970      |     gdp1971      |     gdp1972      |     gdp1973
|     gdp1974      |     gdp1975      |     gdp1976      |     gdp1977      |gdp1978      |     gdp1979
 |     gdp1980      |     gdp1981      |     gdp1982      |     gdp1983      |     gdp1984      |     gdp1985
      |     gdp1986|    gdp1987      |     gdp1988      |     gdp1989      |     gdp1990      |     gdp1991
     |     gdp1992      |     gdp1993      |     gdp1994      |     gdp1995      |     gdp1996      |
gdp1997      |     gdp1998      |     gdp1999      |     gdp2000      |     gdp2001      |     gdp2002
|     gdp2003      |gdp2004      |     gdp2005      |     gdp2006      |     gdp2007      |     gdp2008
  |     gdp2009      |     gdp2010
-------+---------------------+-------------------+-------------------+------------------+------------------
+------------------+-------------------+------------------+------------------+-------------------+---------
----------+-------------------+------------------+-------------------+-------------------+-----------------
--+------------------+-------------------+-------------------+-------------------+------------------+-------
-----------+------------------+-------------------+-------------------+------------------+------------------
-+-------------------+------------------+-------------------+-------------------+-------------------+-------
------------+--------------------+------------------+-------------------+------------------+----------------
---+-------------------+-------------------+------------------+-------------------+------------------+------
------------+------------------+------------------+------------------+------------------+------------------+
------------------+------------------+------------------+------------------+------------------+-------------
-----+------------------+------------------+------------------+------------------+------------------+-------
-----------+------------------+------------------+------------------+------------------+------------------+-
-----------------+------------------+------------------+------------------+------------------+--------------
----+------------------+------------------+------------------+------------------+------------------+--------
----------+------------------+------------------+------------------+------------------+------------------
0.886 | Belgium             |  18585.6613572407 | -16145.6374560074 |  26938.956253415 | 8094.30475779595 |
 12073.5461203817 | -11069.0567600181 | 19133.8584911727|   5500.312894949 | -4227.94863799987 |  6265.77925410752
|  -10884.749295608 | 30929.4669575201 | -7831.49439429977 |  3235.81760508742 | -22765.9285442662 | 27200
.6767714485 | -10554.9550160917 |   1169.4144482273 | -16783.7961289161 | 27932.2660829329 | 17227.9083196848
| 13956.0524012749 | -40175.6286481088 | -10889.4785920499 | 22703.6576872859 | -14635.5832197402 |
2857.12270512168 | 20473.5044214494 | -52199.4895696423 | -11038.7346460738 |  18466.7298633088 | -17410.4225137703 |
-3475.63826305462 | 29305.6753822341 |   1242.5724942049 | 17491.0096310849 | -12609.9984515902 | -17909.3603476248
|  6276.58431412381 | 21851.9475485178 | -2614.33738160397 | 3777.74134131349 | 4522.08854282736 | 4251.90446379366
| 4512.15101396876 | 4265.49424538129 | 5190.06845330997 | 4543.80444817989 | 5639.81122679089 | 4420.44705213467
|  5658.8820279283 | 5172.69025294376 | 5019.63640408663 | 5938.84979495903 | 4976.57073629812 | 4710.49525137591
| 6523.65700286465 | 5067.82520773578 | 6789.13070219317 | 5525.94643553563 | 6894.68336419297 | 5961.58442474331
| 5661.21093840818 | 7721.56088518218 |  5959.7301109143 | 6453.43604137202 | 6739.39384033096 | 7517.97645468455
| 6907.49136910647 | 7049.03921764209 | 7726.49091035527 | 8552.65909911844 | 7963.94487647115 | 7187.45827585515
| 7994.02955410523 | 9532.89844418041 | 7962.25713582666 | 7846.68238907624 | 10230.9878908643 | 8642.76044946519
| 8886.79860331866 |  8718.3731386891
...
(96 rows)

5.4 - APPLY_INVERSE_SVD

Transforms the data back to the original domain.

Transforms the data back to the original domain. This essentially computes the approximated version of the original data by multiplying three matrices: matrix U (input to this function), matrices S and V (stored in the model).

Syntax

APPLY_INVERSE_SVD ( 'input-columns'
        USING PARAMETERS model_name = 'model-name'
            [, exclude_columns = 'excluded-columns']
            [, key_columns = 'key-columns'] )

Arguments

input-columns

Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. The following requirements apply:

All columns must be a numeric data type.
Enclose the column name in double quotes if it contains special characters.

Parameters

model_name: Name of the model (case-insensitive).
exclude_columns: Comma-separated list of column names from input-columns to exclude from processing.
key_columns: Comma-separated list of column names from input-columns that identify its data rows. These columns are included in the output table.

Examples

=> SELECT SVD ('svdmodel', 'small_svd', 'x1,x2,x3,x4');
SVD
--------------------------------------------------------------
Finished in 1 iterations.
Accepted Rows: 8  Rejected Rows: 0
(1 row)

=> CREATE TABLE transform_svd AS SELECT
     APPLY_SVD (id, x1, x2, x3, x4 USING PARAMETERS model_name='svdmodel', exclude_columns='id', key_columns='id')
     OVER () FROM small_svd;
CREATE TABLE

=> SELECT * FROM transform_svd;
id  |       col1        |        col2         |        col3         |        col4
----+-------------------+---------------------+---------------------+--------------------
4   |  0.44849499240202 |  -0.347260956311326 |   0.186958376368345 |  0.378561270493651
6   |  0.17652411036246 | -0.0753183783382909 |  -0.678196192333598 | 0.0567124770173372
1   | 0.494871802886819 |   0.161721379259287 |  0.0712816417153664 | -0.473145877877408
2   |  0.17652411036246 | -0.0753183783382909 |  -0.678196192333598 | 0.0567124770173372
3   | 0.150974762654569 |   0.589561842046029 | 0.00392654610109522 |  0.360011163271921
5   | 0.494871802886819 |   0.161721379259287 |  0.0712816417153664 | -0.473145877877408
8   |  0.44849499240202 |  -0.347260956311326 |   0.186958376368345 |  0.378561270493651
7   | 0.150974762654569 |   0.589561842046029 | 0.00392654610109522 |  0.360011163271921
(8 rows)

=> SELECT APPLY_INVERSE_SVD (* USING PARAMETERS model_name='svdmodel', exclude_columns='id',
key_columns='id') OVER () FROM transform_svd;
id  |        x1        |        x2        |        x3        |        x4
----+------------------+------------------+------------------+------------------
4 | 91.4056627665577 | 44.7629617207482 | 83.1704961993117 | 38.9274292265543
6 | 20.6468626294368 | 9.30974906868751 | 8.71006863405534 |  6.5855928603967
7 | 31.2494347777156 | 20.6336519003026 | 27.5668287751507 | 5.84427645886865
1 |  107.93376580719 | 51.6980548011917 | 97.9665796560552 | 40.4918236881051
2 | 20.6468626294368 | 9.30974906868751 | 8.71006863405534 |  6.5855928603967
3 | 31.2494347777156 | 20.6336519003026 | 27.5668287751507 | 5.84427645886865
5 |  107.93376580719 | 51.6980548011917 | 97.9665796560552 | 40.4918236881051
8 | 91.4056627665577 | 44.7629617207482 | 83.1704961993117 | 38.9274292265543
(8 rows)

5.5 - APPLY_KMEANS

Assigns each row of an input relation to a cluster center from an existing k-means model.

Syntax

APPLY_KMEANS ( input-columns
       USING PARAMETERS model_name = 'model-name' [, match_by_pos = match-by-position] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Privileges

Non-superusers: model owner, or USAGE privileges on the model

Examples

=> SELECT KMEANS('myKmeansModel', 'iris1', '*', 5
USING PARAMETERS max_iterations=20, output_view='myKmeansView', key_columns='id', exclude_columns='Species, id');
           KMEANS
----------------------------
 Finished in 12 iterations

(1 row)
=> SELECT id, APPLY_KMEANS(Sepal_Length, 2.2, 1.3, Petal_Width
USING PARAMETERS model_name='myKmeansModel', match_by_pos='true') FROM iris2;
 id  | APPLY_KMEANS
-----+--------------
   5 |            1
  10 |            1
  14 |            1
  15 |            1
  21 |            1
  22 |            1
  24 |            1
  25 |            1
  32 |            1
  33 |            1
  34 |            1
  35 |            1
  38 |            1
  39 |            1
  42 |            1
...
 (60 rows)

5.6 - APPLY_KPROTOTYPES

Assigns each row of an input relation to a cluster center from an existing k-prototypes model.

Syntax

APPLY_KPROTOTYPES ( input-columns
       USING PARAMETERS model_name = 'model-name' [, match_by_pos = match-by-position] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Privileges

Non-superusers: model owner, or USAGE privileges on the model

Examples

The following example creates k-prototypes model small_model and applies it to input table small_test_mixed:

=> SELECT KPROTOTYPES('small_model_initcenters', 'small_test_mixed', 'x0, country', 3 USING PARAMETERS initial_centers_table='small_test_mixed_centers', key_columns='pid');
      KPROTOTYPES
---------------------------
Finished in 2 iterations

(1 row)

=> SELECT country, x0, APPLY_KPROTOTYPES(country, x0
USING PARAMETERS model_name='small_model')
FROM small_test_mixed;
  country   | x0  | apply_kprototypes
------------+-----+-------------------
 'China'    |  20 |                 0
 'US'       |  85 |                 2
 'Russia'   |  80 |                 1
 'Brazil'   |  78 |                 1
 'US'       |  23 |                 0
 'US'       |  50 |                 0
 'Canada'   |  24 |                 0
 'Canada'   |  18 |                 0
 'Russia'   |  90 |                 2
 'Russia'   |  98 |                 2
 'Brazil'   |  89 |                 2
...
(45 rows)

5.7 - APPLY_NORMALIZE

A UDTF function that applies the normalization parameters saved in a model to a set of specified input columns.

A UDTF function that applies the normalization parameters saved in a model to a set of specified input columns. If any column specified in the function is not in the model, its data passes through unchanged to APPLY_NORMALIZE.

Note

Note: If a column contains only one distinct value, APPLY_NORMALIZE returns NaN for values in that column.

Syntax

APPLY_NORMALIZE ( input-columns USING PARAMETERS model_name = 'model-name');

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns. If you supply an asterisk, APPLY_NORMALIZE normalizes all columns in the model.

Parameters

model_name: Name of the model (case-insensitive).

Examples

The following example creates a model with NORMALIZE_FIT using the wt and hp columns in table mtcars , and then uses this model in successive calls to APPLY_NORMALIZE and REVERSE_NORMALIZE.

=> SELECT NORMALIZE_FIT('mtcars_normfit', 'mtcars', 'wt,hp', 'minmax');
NORMALIZE_FIT
---------------
Success
(1 row)

The following call to APPLY_NORMALIZE specifies the hp and cyl columns in table mtcars, where hp is in the normalization model and cyl is not in the normalization model:

=> CREATE TABLE mtcars_normalized AS SELECT APPLY_NORMALIZE (hp, cyl USING PARAMETERS model_name = 'mtcars_normfit') FROM mtcars;
CREATE TABLE
=> SELECT * FROM mtcars_normalized;
          hp        | cyl
--------------------+-----
  0.434628975265018 | 8
  0.681978798586572 | 8
  0.434628975265018 | 6
                  1 | 8
  0.540636042402827 | 8
                  0 | 4
  0.681978798586572 | 8
 0.0459363957597173 | 4
  0.434628975265018 | 8
  0.204946996466431 | 6
  0.250883392226148 | 6
  0.049469964664311 | 4
  0.204946996466431 | 6
  0.201413427561837 | 4
  0.204946996466431 | 6
  0.250883392226148 | 6
  0.049469964664311 | 4
  0.215547703180212 | 4
 0.0353356890459364 | 4
  0.187279151943463 | 6
  0.452296819787986 | 8
  0.628975265017668 | 8
  0.346289752650177 | 8
  0.137809187279152 | 4
  0.749116607773852 | 8
  0.144876325088339 | 4
  0.151943462897526 | 4
  0.452296819787986 | 8
  0.452296819787986 | 8
  0.575971731448763 | 8
  0.159010600706714 | 4
  0.346289752650177 | 8
(32 rows)

=> SELECT REVERSE_NORMALIZE (hp, cyl USING PARAMETERS model_name='mtcars_normfit') FROM mtcars_normalized;
  hp | cyl
-----+-----
 175 | 8
 245 | 8
 175 | 6
 335 | 8
 205 | 8
  52 | 4
 245 | 8
  65 | 4
 175 | 8
 110 | 6
 123 | 6
  66 | 4
 110 | 6
 109 | 4
 110 | 6
 123 | 6
  66 | 4
 113 | 4
  62 | 4
 105 | 6
 180 | 8
 230 | 8
 150 | 8
  91 | 4
 264 | 8
  93 | 4
  95 | 4
 180 | 8
 180 | 8
 215 | 8
  97 | 4
 150 | 8
(32 rows)

=> SELECT REVERSE_NORMALIZE (hp, cyl USING PARAMETERS model_name='mtcars_normfit') FROM mtcars_normalized;
       hp        | cyl
-----------------+-----
205.000005722046 |   8
150.000000357628 |   8
150.000000357628 |   8
93.0000016987324 |   4
 174.99999666214 |   8
94.9999992102385 |   4
214.999997496605 |   8
97.0000009387732 |   4
245.000006556511 |   8
 174.99999666214 |   6
             335 |   8
245.000006556511 |   8
62.0000002086163 |   4
 174.99999666214 |   8
230.000002026558 |   8
              52 |   4
263.999997675419 |   8
109.999999523163 |   6
123.000002324581 |   6
64.9999996386468 |   4
66.0000005029142 |   4
112.999997898936 |   4
109.999999523163 |   6
180.000000983477 |   8
180.000000983477 |   8
108.999998658895 |   4
109.999999523163 |   6
104.999999418855 |   6
123.000002324581 |   6
180.000000983477 |   8
66.0000005029142 |   4
90.9999999701977 |   4
(32 rows)

5.8 - APPLY_ONE_HOT_ENCODER

A user-defined transform function (UDTF) that loads the one hot encoder model and writes out a table that contains the encoded columns.

Syntax

APPLY_ONE_HOT_ENCODER( input-columns
        USING PARAMETERS model_name = 'model-name'
            [, drop_first = 'is-first']
            [, ignore_null = 'ignore']
            [, separator = 'separator-character']
            [, column_naming = 'name-output']
            [, null_column_name = 'null-column-name'] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

, stores the categories and their corresponding levels.

drop_first

Boolean value, one of the following:

true (default): Treat the first level of the categorical variable as the reference level.
false: Every level of the categorical variable has a corresponding column in the output view

ignore_null

Boolean value, one of the following:

true (default): Null values set all corresponding one-hot binary columns to null.
false: Null values in input-columns are treated as a categorical level

separator

The character that separates the input variable name and the indicator variable level in the output table.To avoid using any separator, set this parameter to null value.

Default: Underscore (_)

column_naming

Appends categorical levels to column names according to the specified method:

indices (default): Uses integer indices to represent categorical levels.
values/values_relaxed: Both methods use categorical level names. If duplicate column names occur, the function attempts to disambiguate them by appending _n, where n is a zero-based integer index (_0, _1,...).

If the function cannot produce unique column names , it handles this according to the chosen method:
- values returns an error.
- values_relaxed reverts to using indices.

Important

The following column naming rules apply if column_naming is set to values or values_relaxed:

Input column names with more than 128 characters are truncated.
Column names can contain special characters.
If parameter ignore_null is set to true, APPLY_ONE_HOT_ENCODER constructs the column name from the value set in parameter null_column_name. If this parameter is omitted, the string null is used.

null_column_name

The string used in naming the indicator column for null values, used only if ignore_null is set to false and column_naming is set to values or values_relaxed.

Default:null

Note

Note: If an input row contains a level not stored in the model, the output row columns corresponding to that categorical level are returned as null values.

Examples

=> SELECT APPLY_ONE_HOT_ENCODER(cyl USING PARAMETERS model_name='one_hot_encoder_model',
drop_first='true', ignore_null='false') FROM mtcars;
cyl | cyl_1 | cyl_2
----+-------+-------
8   |     0 |     1
4   |     0 |     0
4   |     0 |     0
8   |     0 |     1
8   |     0 |     1
8   |     0 |     1
4   |     0 |     0
8   |     0 |     1
8   |     0 |     1
4   |     0 |     0
8   |     0 |     1
6   |     1 |     0
4   |     0 |     0
4   |     0 |     0
6   |     1 |     0
6   |     1 |     0
8   |     0 |     1
8   |     0 |     1
4   |     0 |     0
4   |     0 |     0
6   |     1 |     0
8   |     0 |     1
8   |     0 |     1
6   |     1 |     0
4   |     0 |     0
8   |     0 |     1
8   |     0 |     1
8   |     0 |     1
6   |     1 |     0
6   |     1 |     0
4   |     0 |     0
4   |     0 |     0
(32 rows)

5.9 - APPLY_PCA

Transforms the data using a PCA model.

Transforms the data using a PCA model. This returns new coordinates of each data point.

Syntax

APPLY_PCA ( input-columns
        USING PARAMETERS model_name = 'model-name'
            [, num_components = num-components]
            [, cutoff = cutoff-value]
            [, match_by_pos = match-by-position]
            [, exclude_columns = 'excluded-columns']
            [, key_columns = 'key-columns'] )

Arguments

input-columns

Comma-separated list of columns that contain the data matrix, or asterisk (*) to select all columns. The following requirements apply:

All columns must be a numeric data type.
Enclose the column name in double quotes if it contains special characters.

Parameters

model_name

Name of the model (case-insensitive).

num_components

The number of components to keep in the model. This is the number of output columns that will be generated. If you omit this parameter and the cutoff parameter, all model components are kept.

cutoff

Set to 1, specifies the minimum accumulated explained variance. Components are taken until the accumulated explained variance reaches this value.

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

exclude_columns

Comma-separated list of column names from input-columns to exclude from processing.

key_columns

Comma-separated list of column names from input-columns that identify its data rows. These columns are included in the output table.

Examples


=> SELECT PCA ('pcamodel', 'world','country,HDI,em1970,em1971,em1972,em1973,em1974,em1975,em1976,em1977,
em1978,em1979,em1980,em1981,em1982,em1983,em1984 ,em1985,em1986,em1987,em1988,em1989,em1990,em1991,em1992,
em1993,em1994,em1995,em1996,em1997,em1998,em1999,em2000,em2001,em2002,em2003,em2004,em2005,em2006,em2007,
em2008,em2009,em2010,gdp1970,gdp1971,gdp1972,gdp1973,gdp1974,gdp1975,gdp1976,gdp1977,gdp1978,gdp1979,gdp1980,
gdp1981,gdp1982,gdp1983,gdp1984,gdp1985,gdp1986,gdp1987,gdp1988,gdp1989,gdp1990,gdp1991,gdp1992,gdp1993,
gdp1994,gdp1995,gdp1996,gdp1997,gdp1998,gdp1999,gdp2000,gdp2001,gdp2002,gdp2003,gdp2004,gdp2005,gdp2006,
gdp2007,gdp2008,gdp2009,gdp2010' USING PARAMETERS exclude_columns='HDI,country');
PCA
---------------------------------------------------------------
Finished in 1 iterations.
Accepted Rows: 96  Rejected Rows: 0
(1 row)
=> CREATE TABLE worldPCA AS SELECT
APPLY_PCA (HDI,country,em1970,em1971,em1972,em1973,em1974,em1975,em1976,em1977,em1978,em1979,
em1980,em1981,em1982,em1983,em1984 ,em1985,em1986,em1987,em1988,em1989,em1990,em1991,em1992,em1993,em1994,
em1995,em1996,em1997,em1998,em1999,em2000,em2001,em2002,em2003,em2004,em2005,em2006,em2007,em2008,em2009,
em2010,gdp1970,gdp1971,gdp1972,gdp1973,gdp1974,gdp1975,gdp1976,gdp1977,gdp1978,gdp1979,gdp1980,gdp1981,gdp1982,
gdp1983,gdp1984,gdp1985,gdp1986,gdp1987,gdp1988,gdp1989,gdp1990,gdp1991,gdp1992,gdp1993,gdp1994,gdp1995,
gdp1996,gdp1997,gdp1998,gdp1999,gdp2000,gdp2001,gdp2002,gdp2003,gdp2004,gdp2005,gdp2006,gdp2007,gdp2008,
gdp2009,gdp2010 USING PARAMETERS model_name='pcamodel', exclude_columns='HDI, country', key_columns='HDI,
country',cutoff=.3)OVER () FROM world;
CREATE TABLE

=> SELECT * FROM worldPCA;
HDI   |       country       |       col1
------+---------------------+-------------------
0.886 | Belgium             |  79002.2946705704
0.699 | Belize              | -25631.6670012556
0.427 | Benin               | -40373.4104598122
0.805 | Chile               | -16805.7940082156
0.687 | China               | -37279.2893141103
0.744 | Costa Rica          | -19505.5631231635
0.4   | Cote d'Ivoire       | -38058.2060339272
0.776 | Cuba                | -23724.5779612041
0.895 | Denmark             |  117325.594028813
0.644 | Egypt               | -34609.9941604549
...
(96 rows)

=> SELECT APPLY_INVERSE_PCA (HDI, country, col1
    USING PARAMETERS model_name = 'pcamodel', exclude_columns='HDI,country',
    key_columns = 'HDI, country') OVER () FROM worldPCA;
HDI  |       country       |      em1970       |      em1971       |      em1972      |      em1973      |
      em1974      |      em1975       |      em1976|      em1977      |      em1978       |      em1979
   |      em1980       |      em1981      |      em1982       |      em1983       |      em1984       |em1985
|      em1986       |      em1987       |      em1988       |      em1989      |      em1990      |      em1991
|      em1992       |      em1993|      em1994      |      em1995       |      em1996       |      em1997
    |      em1998       |      em1999       |      em2000       |      em2001       |em2002       |
em2003      |      em2004       |      em2005      |      em2006       |      em2007       |      em2008
|      em2009      |      em2010       |     gdp1970      |     gdp1971      |     gdp1972      |     gdp1973
|     gdp1974      |     gdp1975      |     gdp1976      |     gdp1977      |gdp1978      |     gdp1979
 |     gdp1980      |     gdp1981      |     gdp1982      |     gdp1983      |     gdp1984      |     gdp1985
      |     gdp1986|    gdp1987      |     gdp1988      |     gdp1989      |     gdp1990      |     gdp1991
     |     gdp1992      |     gdp1993      |     gdp1994      |     gdp1995      |     gdp1996      |
gdp1997      |     gdp1998      |     gdp1999      |     gdp2000      |     gdp2001      |     gdp2002
|     gdp2003      |gdp2004      |     gdp2005      |     gdp2006      |     gdp2007      |     gdp2008
  |     gdp2009      |     gdp2010
-------+---------------------+-------------------+-------------------+------------------+------------------
+------------------+-------------------+------------------+------------------+-------------------+---------
----------+-------------------+------------------+-------------------+-------------------+-----------------
--+------------------+-------------------+-------------------+-------------------+------------------+-------
-----------+------------------+-------------------+-------------------+------------------+------------------
-+-------------------+------------------+-------------------+-------------------+-------------------+-------
------------+--------------------+------------------+-------------------+------------------+----------------
---+-------------------+-------------------+------------------+-------------------+------------------+------
------------+------------------+------------------+------------------+------------------+------------------+
------------------+------------------+------------------+------------------+------------------+-------------
-----+------------------+------------------+------------------+------------------+------------------+-------
-----------+------------------+------------------+------------------+------------------+------------------+-
-----------------+------------------+------------------+------------------+------------------+--------------
----+------------------+------------------+------------------+------------------+------------------+--------
----------+------------------+------------------+------------------+------------------+------------------
0.886 | Belgium             |  18585.6613572407 | -16145.6374560074 |  26938.956253415 | 8094.30475779595 |
 12073.5461203817 | -11069.0567600181 | 19133.8584911727|   5500.312894949 | -4227.94863799987 |  6265.77925410752
|  -10884.749295608 | 30929.4669575201 | -7831.49439429977 |  3235.81760508742 | -22765.9285442662 | 27200
.6767714485 | -10554.9550160917 |   1169.4144482273 | -16783.7961289161 | 27932.2660829329 | 17227.9083196848
| 13956.0524012749 | -40175.6286481088 | -10889.4785920499 | 22703.6576872859 | -14635.5832197402 |
2857.12270512168 | 20473.5044214494 | -52199.4895696423 | -11038.7346460738 |  18466.7298633088 | -17410.4225137703 |
-3475.63826305462 | 29305.6753822341 |   1242.5724942049 | 17491.0096310849 | -12609.9984515902 | -17909.3603476248
|  6276.58431412381 | 21851.9475485178 | -2614.33738160397 | 3777.74134131349 | 4522.08854282736 | 4251.90446379366
| 4512.15101396876 | 4265.49424538129 | 5190.06845330997 | 4543.80444817989 | 5639.81122679089 | 4420.44705213467
|  5658.8820279283 | 5172.69025294376 | 5019.63640408663 | 5938.84979495903 | 4976.57073629812 | 4710.49525137591
| 6523.65700286465 | 5067.82520773578 | 6789.13070219317 | 5525.94643553563 | 6894.68336419297 | 5961.58442474331
| 5661.21093840818 | 7721.56088518218 |  5959.7301109143 | 6453.43604137202 | 6739.39384033096 | 7517.97645468455
| 6907.49136910647 | 7049.03921764209 | 7726.49091035527 | 8552.65909911844 | 7963.94487647115 | 7187.45827585515
| 7994.02955410523 | 9532.89844418041 | 7962.25713582666 | 7846.68238907624 | 10230.9878908643 | 8642.76044946519
| 8886.79860331866 |  8718.3731386891
...
(96 rows)

5.10 - APPLY_SVD

Transforms the data using an SVD model.

Transforms the data using an SVD model. This computes the matrix U of the SVD decomposition.

Syntax

APPLY_SVD ( input-columns
        USING PARAMETERS model_name = 'model-name'
            [, num_components = num-components]
            [, cutoff = cutoff-value]
            [, match_by_pos = match-by-position]
            [, exclude_columns = 'excluded-columns']
            [, key_columns = 'key-columns'] )

Arguments

input-columns

Comma-separated list of columns that contain the data matrix, or asterisk (*) to select all columns. The following requirements apply:

All columns must be a numeric data type.
Enclose the column name in double quotes if it contains special characters.

Parameters

model_name

Name of the model (case-insensitive).

num_components

The number of components to keep in the model. This is the number of output columns that will be generated. If neither this parameter nor the cutoff parameter is provided, all components from the model are kept.

cutoff

Set to 1, specifies the minimum accumulated explained variance. Components are taken until the accumulated explained variance reaches this value. If you omit this parameter and the num_components parameter, all model components are kept.

match_by_pos

Boolean value that specifies how input columns are matched to model columns:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

exclude_columns

Comma-separated list of column names from input-columns to exclude from processing.

key_columns

Comma-separated list of column names from input-columns that identify its data rows. These columns are included in the output table.

Examples

=> SELECT SVD ('svdmodel', 'small_svd', 'x1,x2,x3,x4');
SVD
--------------------------------------------------------------
Finished in 1 iterations.
Accepted Rows: 8  Rejected Rows: 0
(1 row)

=> CREATE TABLE transform_svd AS SELECT
     APPLY_SVD (id, x1, x2, x3, x4 USING PARAMETERS model_name='svdmodel', exclude_columns='id', key_columns='id')
     OVER () FROM small_svd;
CREATE TABLE

=> SELECT * FROM transform_svd;
id  |       col1        |        col2         |        col3         |        col4
----+-------------------+---------------------+---------------------+--------------------
4   |  0.44849499240202 |  -0.347260956311326 |   0.186958376368345 |  0.378561270493651
6   |  0.17652411036246 | -0.0753183783382909 |  -0.678196192333598 | 0.0567124770173372
1   | 0.494871802886819 |   0.161721379259287 |  0.0712816417153664 | -0.473145877877408
2   |  0.17652411036246 | -0.0753183783382909 |  -0.678196192333598 | 0.0567124770173372
3   | 0.150974762654569 |   0.589561842046029 | 0.00392654610109522 |  0.360011163271921
5   | 0.494871802886819 |   0.161721379259287 |  0.0712816417153664 | -0.473145877877408
8   |  0.44849499240202 |  -0.347260956311326 |   0.186958376368345 |  0.378561270493651
7   | 0.150974762654569 |   0.589561842046029 | 0.00392654610109522 |  0.360011163271921
(8 rows)

=> SELECT APPLY_INVERSE_SVD (* USING PARAMETERS model_name='svdmodel', exclude_columns='id',
key_columns='id') OVER () FROM transform_svd;
id  |        x1        |        x2        |        x3        |        x4
----+------------------+------------------+------------------+------------------
4 | 91.4056627665577 | 44.7629617207482 | 83.1704961993117 | 38.9274292265543
6 | 20.6468626294368 | 9.30974906868751 | 8.71006863405534 |  6.5855928603967
7 | 31.2494347777156 | 20.6336519003026 | 27.5668287751507 | 5.84427645886865
1 |  107.93376580719 | 51.6980548011917 | 97.9665796560552 | 40.4918236881051
2 | 20.6468626294368 | 9.30974906868751 | 8.71006863405534 |  6.5855928603967
3 | 31.2494347777156 | 20.6336519003026 | 27.5668287751507 | 5.84427645886865
5 |  107.93376580719 | 51.6980548011917 | 97.9665796560552 | 40.4918236881051
8 | 91.4056627665577 | 44.7629617207482 | 83.1704961993117 | 38.9274292265543
(8 rows)

5.11 - PREDICT_ARIMA

Applies an autoregressive integrated moving average (ARIMA) model to an input relation or makes predictions using the in-sample data.

Applies an autoregressive integrated moving average (ARIMA) model to an input relation or makes predictions using the in-sample data. ARIMA models make predictions based on preceding time series values and errors of previous predictions. The function, by default, returns the predicted values plus the mean of the model.

Behavior type

Immutable

Syntax

Apply to an input relation:

PREDICT_ARIMA ( timeseries-column
        USING PARAMETERS param=value[,...] )
        OVER (ORDER BY timestamp-column)
        FROM input-relation

Make predictions using the in-sample data:

PREDICT_ARIMA ( USING PARAMETERS model_name = 'ARIMA-model'
        [, start = prediction-start ]
        [, npredictions = num-predictions ]
        [, output_standard_errors = boolean ] )
        OVER ()

Arguments

timeseries-column: Name of a NUMERIC column in input-relation used to make predictions.
timestamp-column: Name of an INTEGER, FLOAT, or TIMESTAMP column in input-relation that represents the timestamp variable. The timestep between consecutive entries should be consistent throughout the timestamp-column.
input-relation: Input relation containing timeseries-column and timestamp-column.

Parameters

model_name

Name of a trained ARIMA model.

start

The behavior of the start parameter and its range of accepted values depends on whether you provide a timeseries-column:

No provided timeseries-column: start must be an integer ≥0, where zero indicates to start prediction at the end of the in-sample data. If start is a positive value, the function predicts the values between the end of the in-sample data and the start index, and then uses the predicted values as time series inputs for the subsequent npredictions.
timeseries-column provided: start must be an integer ≥1 and identifies the index (row) of the timeseries-column at which to begin prediction. If the start index is greater than the number of rows, N, in the input data, the function predicts the values between N and start and uses the predicted values as time series inputs for the subsequent npredictions.

Default:

No provided timeseries-column: prediction begins from the end of the in-sample data.
timeseries-column provided: prediction begins from the end of the provided input data.

npredictions

Integer ≥1, the number of predicted timesteps.

Default: 10

missing

Methods for handling missing values, one of the following strings:

'drop': Missing values are ignored.
'error': Missing values raise an error.
'zero': Missing values are replaced with 0.
'linear_interpolation': Missing values are replaced by linearly-interpolated values based on the nearest valid entries before and after the missing value. If all values before or after a missing value in the prediction range are missing or invalid, interpolation is impossible and the function errors.

Default: Method used when training the model

add_mean

Boolean, whether to add the model mean to the predicted value.

Default: True

output_standard_errors

Boolean, whether to return estimates of the standard error of each prediction.

Default: False

Examples

The following example makes predictions using the in-sample data that the arima_temp model was trained on:

=> SELECT PREDICT_ARIMA(USING PARAMETERS model_name='arima_temp', npredictions=10) OVER();
   prediction
------------------
12.9797364979873
13.3768377212635
13.460660717892
13.468204126011
13.4572461558472
13.4418721036084
13.425515187182
13.4090117135945
13.3925648829068
13.3762235523779
(10 rows)

You can also apply the model to an input relation:

=> SELECT PREDICT_ARIMA(temperature USING PARAMETERS model_name='arima_temp', start=100, npredictions=10) OVER(ORDER BY time) FROM temp_data;
   prediction
------------------
15.0373229398431
13.4709102391534
10.5720766977885
13.1971253722069
13.5615497506689
13.1613971089657
13.4008120147841
12.612020423044
12.9026197179173
13.2392824099367
(10 rows)

For an in-depth example that trains and makes predictions with an ARIMA model, see ARIMA model example.

5.12 - PREDICT_AUTOREGRESSOR

Applies an autoregressor (AR) model to an input relation.

Autoregressive models use previous values to make predictions. More specifically, the user-specified "lag" determines how many previous timesteps it takes into account during computation, and predicted values are linear combinations of those lags.

Syntax

PREDICT_AUTOREGRESSOR ( timeseries-column
        USING PARAMETERS
            model-name = 'model-name'
            [, start = starting-index]
            [, npredictions = npredictions]
            [, missing = "imputation-method" ] )
        OVER (ORDER BY timestamp-column)
        FROM input-relation

Note

The following argument, as written, is required and cannot be omitted nor substituted with another type of clause.

OVER (ORDER BY timestamp-column)

Arguments

timeseries-column: The timeseries column used to make the prediction (only the last p values, specified during model creation, are used).
timestamp-column: The timestamp column, with consistent timesteps, used to make the prediction.
input-relation: The input relation containing the timeseries-column and timestamp-column.
Note that input-relation cannot have missing values in any of the p (set during training) rows preceding start. To handle missing values, see IMPUTE or Linear interpolation.

Parameters

model_name

Name of the model (case-insensitive).

start

INTEGER >p or ≤0, the index (row) of the input-relation at which to start the prediction. If omitted, the prediction starts at the end of the input-relation.

If the start index is greater than the number of rows N in timeseries-column, then the values between N and start are predicted and used for the prediction.

If negative, the start index is identified by counting backwards from the end of the input-relation.

For an input-relation of N rows, negative values have a lower limit of either -1000 or -(N-p), whichever is greater.

Default: the end of input-relation

npredictions

INTEGER ≥1, the number of predicted timesteps.

Default: 10

missing

One of the following methods for handling missing values:

drop: Missing values are ignored.
error: Missing values raise an error.
zero: Missing values are replaced with 0.
linear_interpolation: Missing values are replaced by linearly-interpolated values based on the nearest valid entries before and after the missing value. If all values before or after a missing value in the prediction range are missing or invalid, interpolation is impossible and the function errors.

Default: Method used when training the model

Examples

See Autoregressive model example.

5.13 - PREDICT_LINEAR_REG

Applies a linear regression model on an input relation and returns the predicted value as a FLOAT.

Syntax

PREDICT_LINEAR_REG ( input-columns
        USING PARAMETERS model_name = 'model-name' [, match_by_pos = match-by-position] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Examples

=> SELECT PREDICT_LINEAR_REG(waiting USING PARAMETERS model_name='myLinearRegModel')FROM
faithful ORDER BY id;


 PREDICT_LINEAR_REG
--------------------
   4.15403481386324
   2.18505296804024
   3.76023844469864
    2.8151271587036
   4.62659045686076
   2.26381224187316
   4.86286827835952
   4.62659045686076
   1.94877514654148
   4.62659045686076
   2.18505296804024
...
 (272 rows)

The following example shows how to use the PREDICT_LINEAR_REG function on an input table, using the match_by_pos parameter. Note that you can replace the column argument with a constant that does not match an input column:

=> SELECT PREDICT_LINEAR_REG(55 USING PARAMETERS model_name='linear_reg_faithful',
                     match_by_pos='true')FROM faithful ORDER BY id;
 PREDICT_LINEAR_REG
--------------------
2.28552115094171
2.28552115094171
2.28552115094171
2.28552115094171
2.28552115094171
2.28552115094171
2.28552115094171
...
 (272 rows)

5.14 - PREDICT_LOGISTIC_REG

Applies a logistic regression model on an input relation.

PREDICT_LOGISTIC_REG returns as a FLOAT the predicted class or the probability of the predicted class, depending on how the type parameter is set. You can cast the return value to INTEGER or another numeric type when the return is in the probability of the predicted class.

Syntax

PREDICT_LOGISTIC_REG ( input-columns
        USING PARAMETERS model_name = 'model-name'
            [, type = 'prediction-type']
            [, cutoff = probability-cutoff]
            [, match_by_pos = match-by-position] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

type

Type of prediction for logistic regression, one of the following:

response (default): Predicted values are 0 or 1.
probability: Output is the probability of the predicted category to be 1.

cutoff

Used in conjunction with the type parameter, a FLOAT between 0 and 1, exclusive. When type is set to response, the returned value of prediction is 1 if its corresponding probability is greater than or equal to the value of cutoff; otherwise, it is 0.

Default: 0.5

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Examples

=> SELECT car_model,
            PREDICT_LOGISTIC_REG(mpg, cyl, disp, drat, wt, qsec, vs, gear, carb
                                 USING PARAMETERS model_name='myLogisticRegModel')
          FROM mtcars;
      car_model      | PREDICT_LOGISTIC_REG
---------------------+----------------------
 Camaro Z28          |                    0
 Fiat 128            |                    1
 Fiat X1-9           |                    1
 Ford Pantera L      |                    1
 Merc 450SE          |                    0
 Merc 450SL          |                    0
 Toyota Corona       |                    0
 AMC Javelin         |                    0
 Cadillac Fleetwood  |                    0
 Datsun 710          |                    1
 Dodge Challenger    |                    0
 Hornet 4 Drive      |                    0
 Lotus Europa        |                    1
 Merc 230            |                    0
 Merc 280            |                    0
 Merc 280C           |                    0
 Merc 450SLC         |                    0
 Pontiac Firebird    |                    0
 Porsche 914-2       |                    1
 Toyota Corolla      |                    1
 Valiant             |                    0
 Chrysler Imperial   |                    0
 Duster 360          |                    0
 Ferrari Dino        |                    1
 Honda Civic         |                    1
 Hornet Sportabout   |                    0
 Lincoln Continental |                    0
 Maserati Bora       |                    1
 Mazda RX4           |                    1
 Mazda RX4 Wag       |                    1
 Merc 240D           |                    0
 Volvo 142E          |                    1
(32 rows)

The following example shows how to use PREDICT_LOGISTIC_REG on an input table, using the match_by_pos parameter. Note that you can replace any of the column inputs with a constant that does not match an input column. In this example, column mpg was replaced with the constant 20:

=> SELECT car_model,
            PREDICT_LOGISTIC_REG(20, cyl, disp, drat, wt, qsec, vs, gear, carb
                                 USING PARAMETERS model_name='myLogisticRegModel', match_by_pos='true')
          FROM mtcars;
      car_model     | PREDICT_LOGISTIC_REG
--------------------+----------------------
AMC Javelin         |                    0
Cadillac Fleetwood  |                    0
Camaro Z28          |                    0
Chrysler Imperial   |                    0
Datsun 710          |                    1
Dodge Challenger    |                    0
Duster 360          |                    0
Ferrari Dino        |                    1
Fiat 128            |                    1
Fiat X1-9           |                    1
Ford Pantera L      |                    1
Honda Civic         |                    1
Hornet 4 Drive      |                    0
Hornet Sportabout   |                    0
Lincoln Continental |                    0
Lotus Europa        |                    1
Maserati Bora       |                    1
Mazda RX4           |                    1
Mazda RX4 Wag       |                    1
Merc 230            |                    0
Merc 240D           |                    0
Merc 280            |                    0
Merc 280C           |                    0
Merc 450SE          |                    0
Merc 450SL          |                    0
Merc 450SLC         |                    0
Pontiac Firebird    |                    0
Porsche 914-2       |                    1
Toyota Corolla      |                    1
Toyota Corona       |                    0
Valiant             |                    0
Volvo 142E          |                    1
(32 rows)

5.15 - PREDICT_MOVING_AVERAGE

Applies a moving-average (MA) model, created by MOVING_AVERAGE, to an input relation.

Moving average models use the errors of previous predictions to make future predictions. More specifically, the user-specified "lag" determines how many previous predictions and errors it takes into account during computation.

Syntax

PREDICT_MOVING_AVERAGE ( timeseries-column
        USING PARAMETERS
            model_name = 'model-name'
            [, start = starting-index]
            [, npredictions = npredictions]
            [, missing = "imputation-method" ] )
        OVER (ORDER BY timestamp-column)
        FROM input-relation

Note

The following argument, as written, is required and cannot be omitted nor substituted with another type of clause.

OVER (ORDER BY timestamp-column)

Arguments

timeseries-column: The timeseries column used to make the prediction (only the last q values, specified during model creation, are used).
timestamp-column: The timestamp column, with consistent timesteps, used to make the prediction.
input-relation: The input relation containing the timeseries-column and timestamp-column.
Note that input-relation cannot have missing values in any of the q (set during training) rows preceding start. To handle missing values, see IMPUTE or Linear interpolation.

Parameters

model_name

Name of the model (case-insensitive).

start

INTEGER >q or ≤0, the index (row) of the input-relation at which to start the prediction. If omitted, the prediction starts at the end of the input-relation.

If the start index is greater than the number of rows N in timeseries-column, then the values between N and start are predicted and used for the prediction.

If negative, the start index is identified by counting backwards from the end of the input-relation.

For an input-relation of N rows, negative values have a lower limit of either -1000 or -(N-q), whichever is greater.

Default: the end of input-relation

npredictions

INTEGER ≥1, the number of predicted timesteps.

Default: 10

missing

One of the following methods for handling missing values:

drop: Missing values are ignored.
error: Missing values raise an error.
zero: Missing values are replaced with 0.
linear_interpolation: Missing values are replaced by linearly-interpolated values based on the nearest valid entries before and after the missing value. If all values before or after a missing value in the prediction range are missing or invalid, interpolation is impossible and the function errors.

Default: Method used when training the model

Examples

See Moving-average model example.

5.16 - PREDICT_NAIVE_BAYES

Applies a Naive Bayes model on an input relation.

Depending on how the type parameter is set, PREDICT_NAIVE_BAYES returns a VARCHAR that specifies either the predicted class or probability of the predicted class. If the function returns probability, you can cast the return value to an INTEGER or another numeric data type.

Syntax

PREDICT_NAIVE_BAYES ( input-columns
        USING PARAMETERS model_name = 'model-name'
            [, type = ' return-type ']
            [, class = 'user-input-class']
            [, match_by_pos = match-by-position] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

type

One of the following:

response (default): Returns the class with the highest probability.
probability: Valid only if class parameter is set, returns the probability of belonging to the specified class argument.

class

Required if type parameter is set to probability. If you omit this parameter, PREDICT_NAIVE_BAYES returns the class that it predicts as having the highest probability.

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Examples

=> SELECT party, PREDICT_NAIVE_BAYES (vote1, vote2, vote3
                                        USING PARAMETERS model_name='naive_house84_model',
                                                         type='response')
                                        AS Predicted_Party
                                        FROM house84_test;
   party    | Predicted_Party
------------+-----------------
 democrat   | democrat
 democrat   | democrat
 democrat   | democrat
 republican | republican
 democrat   | democrat
 democrat   | democrat
 democrat   | democrat
 democrat   | democrat
 democrat   | democrat
 republican | republican
 democrat   | democrat
 democrat   | democrat
 democrat   | democrat
 democrat   | republican
 republican | republican
 democrat   | democrat
 republican | republican
...
(99 rows)

5.17 - PREDICT_NAIVE_BAYES_CLASSES

Applies a Naive Bayes model on an input relation and returns the probabilities of classes:.

Applies a Naive Bayes model on an input relation and returns the probabilities of classes:

VARCHAR predicted column contains the class label with the highest probability.
Multiple FLOAT columns, where the first probability column contains the probability for the class specified in the predicted column. Other columns contain the probability of belonging to each class specified in the classes parameter.

Syntax

PREDICT_NAIVE_BAYES_CLASSES ( predictor-columns
        USING PARAMETERS model_name = 'model-name'
            [, key_columns = 'key-columns']
            [, exclude_columns = 'excluded-columns]
            [, classes = 'classes']
            [, match_by_pos = match-by-position] )
OVER( [window-partition-clause] )

Arguments

predictor-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

key_columns

Comma-separated list of predictor column names that identify the output rows. To exclude these and other predictor columns from being used for prediction, include them in the argument list for parameter exclude_columns.

exclude_columns

Comma-separated list of columns from predictor-columns to exclude from processing.

classes

Comma-separated list of class labels in the model. The probability of belonging to this given class as predicted by the classifier. The values are case sensitive.

match_by_pos

Boolean value that specifies how predictor columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the predictor columns list.

Examples

=> SELECT PREDICT_NAIVE_BAYES_CLASSES (id, vote1, vote2 USING PARAMETERS
model_name='naive_house84_model',key_columns='id',exclude_columns='id',
classes='democrat, republican', match_by_pos='false')
        OVER() FROM house84_test;
 id  | Predicted  |    Probability    |     democrat      |    republican
-----+------------+-------------------+-------------------+-------------------
  21 | democrat   | 0.775473383353576 | 0.775473383353576 | 0.224526616646424
  28 | democrat   | 0.775473383353576 | 0.775473383353576 | 0.224526616646424
  83 | republican | 0.592510497724379 | 0.407489502275621 | 0.592510497724379
 102 | democrat   | 0.779889432167111 | 0.779889432167111 | 0.220110567832889
 107 | republican | 0.598662714551597 | 0.401337285448403 | 0.598662714551597
 125 | republican | 0.598662714551597 | 0.401337285448403 | 0.598662714551597
 132 | republican | 0.592510497724379 | 0.407489502275621 | 0.592510497724379
 136 | republican | 0.592510497724379 | 0.407489502275621 | 0.592510497724379
 155 | republican | 0.598662714551597 | 0.401337285448403 | 0.598662714551597
 174 | republican | 0.592510497724379 | 0.407489502275621 | 0.592510497724379
...
(1 row)

5.18 - PREDICT_PMML

Applies an imported PMML model on an input relation.

Applies an imported PMML model on an input relation. The function returns the result that would be expected for the model type encoded in the PMML model.

PREDICT_PMML returns NULL in the following cases:

The predictor is an invalid or NULL value.
The categorical predictor is of an unknown class.

Note

PREDICT_PMML returns values of complex type ROW for models that use the Output tag. Currently, Vertica does not support directly inserting this data into a table.

You can work around this limitation by changing the output to JSON with TO_JSON before inserting it into a table:

=> CREATE TABLE predicted_output AS SELECT TO_JSON(PREDICT_PMML(X1,X2,X3
USING PARAMETERS model_name='pmml_imported_model'))
AS predicted_value
FROM input_table;

Syntax

PREDICT_PMML ( input-columns
        USING PARAMETERS model_name = 'model-name' [, match_by_pos = match-by-position] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive). For a list of supported PMML model types and tags, see PMML features and attributes.

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Examples

In this example, the function call uses all the columns from the table as predictors and predicts the value using the 'my_kmeans' model in PMML format:

SELECT PREDICT_PMML(* USING PARAMETERS model_name='my_kmeans') AS predicted_label FROM table;

In this example, the function call takes only columns col1, col2 as predictors, and predicts the value for each row using the 'my_kmeans' model from schema 'my_schema':

SELECT PREDICT_PMML(col1, col2 USING PARAMETERS model_name='my_schema.my_kmeans') AS predicted_label FROM table;

In this example, the function call returns an error as neither schema nor model-name can accept * as a value:

SELECT PREDICT_PMML(* USING PARAMETERS model_name='*.*') AS predicted_label FROM table;
SELECT PREDICT_PMML(* USING PARAMETERS model_name='*') AS predicted_label FROM table;
SELECT PREDICT_PMML(* USING PARAMETERS model_name='models.*') AS predicted_label FROM table;

5.19 - PREDICT_POISSON_REG

Applies a Poisson regression model on an input relation and returns the predicted value as a FLOAT.

Syntax

PREDICT_POISSON_REG ( input-columns
        USING PARAMETERS model_name = 'model-name' [, match_by_pos = match-by-position] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Examples

=> SELECT PREDICT_POISSON_REG(waiting USING PARAMETERS model_name='MYModel')::numeric(20,10) FROM lin.faithful ORDER BY id;
predict_poisson_reg
---------------------
4.0230080811
2.2284857176
3.5747254723
2.6921731651
4.6357580051
2.2817680621
4.9762900161
4.6357580051
2.0759884314
(9 rows)

5.20 - PREDICT_RF_CLASSIFIER

Applies a random forest model on an input relation.

Applies a random forest model on an input relation. PREDICT_RF_CLASSIFIER returns a VARCHAR data type that specifies one of the following, as determined by how the type parameter is set:

The predicted class (based on popular votes)
Probability of a class for each input instance.

Note

The predicted class is selected only based on the popular vote of the decision trees in the forest. Therefore, in special cases the calculated probability of the predicted class may not be the highest.

Syntax

PREDICT_RF_CLASSIFIER ( input-columns
        USING PARAMETERS model_name = 'model-name'
            [, type = 'prediction-type']
            [, class = 'user-input-class']
            [, match_by_pos = match-by-position] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

type

Type of prediction to return, one of the following:

response (default): The class with the highest probability among all possible classes.
probability: Valid only if the class parameter is set, returns the probability of the specified class.

class

Class to use when the type parameter is set to probability. If you omit this parameter, the function uses the predicted class—the one with the popular vote. Thus, the predict function returns the probability that the input instance belongs to its predicted class.

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Examples

=> SELECT PREDICT_RF_CLASSIFIER (Sepal_Length, Sepal_Width, Petal_Length, Petal_Width
       USING PARAMETERS model_name='myRFModel') FROM iris;
PREDICT_RF_CLASSIFIER
-----------------------
setosa
setosa
setosa
...
versicolor
versicolor
versicolor
...
virginica
virginica
virginica
...
(150 rows)

This example shows how you can use the PREDICT_RF_CLASSIFIER function, using the match_by_pos parameter:

=> SELECT PREDICT_RF_CLASSIFIER (Sepal_Length, Sepal_Width, Petal_Length, Petal_Width
                                USING PARAMETERS model_name='myRFModel', match_by_pos='true') FROM iris;
PREDICT_RF_CLASSIFIER
-----------------------
setosa
setosa
setosa
...
versicolor
versicolor
versicolor
...
virginica
virginica
virginica
...
(150 rows)

5.21 - PREDICT_RF_CLASSIFIER_CLASSES

Applies a random forest model on an input relation and returns the probabilities of classes:.

Applies a random forest model on an input relation and returns the probabilities of classes:

VARCHAR predicted column contains the class label with the highest vote (popular vote).
Multiple FLOAT columns, where the first probability column contains the probability for the class reported in the predicted column. Other columns contain the probability of each class specified in the classes parameter.
Key columns with the same value and data type as matching input columns specified in parameter key_columns.

Note

Selection of the predicted class is based on the popular vote of decision trees in the forest. Thus, in special cases the calculated probability of the predicted class might not be the highest.

Syntax

PREDICT_RF_CLASSIFIER_CLASSES ( predictor-columns
        USING PARAMETERS model_name = 'model-name'
            [, key_columns = 'key-columns']
            [, exclude_columns = 'excluded-columns']
            [, classes = 'classes']
            [, match_by_pos = match-by-position] )
OVER( [window-partition-clause] )

Arguments

predictor-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

key_columns

exclude_columns

Comma-separated list of columns from predictor-columns to exclude from processing.

classes

Comma-separated list of class labels in the model. The probability of belonging to this given class is predicted by the classifier. Values are case sensitive.

match_by_pos

Boolean value that specifies how predictor columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the predictor columns list.

Examples

=> SELECT PREDICT_RF_CLASSIFIER_CLASSES(Sepal_Length, Sepal_Width, Petal_Length, Petal_Width
                               USING PARAMETERS model_name='myRFModel') OVER () FROM iris;
predicted  |    probability
-----------+-------------------
setosa     |                 1
setosa     |              0.99
setosa     |                 1
setosa     |                 1
setosa     |                 1
setosa     |              0.97
setosa     |                 1
setosa     |                 1
setosa     |                 1
setosa     |                 1
setosa     |              0.99
...
(150 rows)

This example shows how to use function PREDICT_RF_CLASSIFIER_CLASSES, using the match_by_pos parameter:

=> SELECT PREDICT_RF_CLASSIFIER_CLASSES(Sepal_Length, Sepal_Width, Petal_Length, Petal_Width
                          USING PARAMETERS model_name='myRFModel', match_by_pos='true') OVER () FROM iris;
predicted  |    probability
-----------+-------------------
setosa     |                 1
setosa     |                 1
setosa     |                 1
setosa     |                 1
setosa     |                 1
setosa     |                 1
setosa     |                 1
setosa     |                 1
setosa     |                 1
setosa     |                 1
setosa     |                 1
...
(150 rows)s

5.22 - PREDICT_RF_REGRESSOR

Applies a random forest model on an input relation, and returns with a FLOAT data type that specifies the predicted value of the random forest model—the average of the prediction of the trees in the forest.

Syntax

PREDICT_RF_REGRESSOR ( input-columns
        USING PARAMETERS model_name = 'model-name' [, match_by_pos = match-by-position] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Examples

=> SELECT PREDICT_RF_REGRESSOR (mpg,cyl,hp,drat,wt
USING PARAMETERS model_name='myRFRegressorModel')FROM mtcars;
PREDICT_RF_REGRESSOR
----------------------
2.94774203574204
2.6954087024087
2.6954087024087
2.89906346431346
2.97688489288489
2.97688489288489
2.7086587024087
2.92078965478965
2.97688489288489
2.7086587024087
2.95621822621823
2.82255155955156
2.7086587024087
2.7086587024087
2.85650394050394
2.85650394050394
2.97688489288489
2.95621822621823
2.6954087024087
2.6954087024087
2.84493251193251
2.97688489288489
2.97688489288489
2.8856467976468
2.6954087024087
2.92078965478965
2.97688489288489
2.97688489288489
2.7934087024087
2.7934087024087
2.7086587024087
2.72469441669442
(32 rows)

5.23 - PREDICT_SVM_CLASSIFIER

Uses an SVM model to predict class labels for samples in an input relation, and returns the predicted value as a FLOAT data type.

Syntax

PREDICT_SVM_CLASSIFIER (input-columns
     USING PARAMETERS model_name = 'model-name'
            [, match_by_pos = match-by-position]
            [, type = 'return-type']
            [, cutoff = 'cutoff-value'] ] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

type

A string that specifies the output to return for each input row, one of the following:

response: Outputs the predicted class of 0 or 1.
probability: Outputs a value in the range (0,1), the prediction score transformed using the logistic function.

cutoff

Valid only if the type parameter is set to probability, a FLOAT value that is compared to the transformed prediction score to determine the predicted class.

Default: 0

Examples


=> SELECT PREDICT_SVM_CLASSIFIER (mpg,cyl,disp,wt,qsec,vs,gear,carb
USING PARAMETERS model_name='mySvmClassModel') FROM mtcars;
PREDICT_SVM_CLASSIFIER
------------------------
0
0
1
0
0
1
1
1
1
0
0
1
0
0
1
0
0
0
0
0
0
1
1
0
0
1
1
1
1
0
0
0
(32 rows)

This example shows how to use PREDICT_SVM_CLASSIFIER on the mtcars table, using the match_by_pos parameter. In this example, column mpg was replaced with the constant 40:


=> SELECT PREDICT_SVM_CLASSIFIER (40,cyl,disp,wt,qsec,vs,gear,carb
USING PARAMETERS model_name='mySvmClassModel', match_by_pos ='true') FROM mtcars;
PREDICT_SVM_CLASSIFIER
------------------------
0
0
0
0
1
0
0
1
1
1
1
1
0
0
0
1
1
1
1
0
0
0
0
0
0
0
0
1
1
0
0
1
(32 rows)

5.24 - PREDICT_SVM_REGRESSOR

Uses an SVM model to perform regression on samples in an input relation, and returns the predicted value as a FLOAT data type.

Syntax

PREDICT_SVM_REGRESSOR(input-columns
        USING PARAMETERS model_name = 'model-name' [, match_by_pos = match-by-position] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Examples


=> SELECT PREDICT_SVM_REGRESSOR(waiting USING PARAMETERS model_name='mySvmRegModel')
           FROM faithful ORDER BY id;
 PREDICT_SVM_REGRESSOR
--------------------
   4.06488248694445
   2.30392277646291
   3.71269054484815
   2.867429883817
   4.48751281746003
   2.37436116488217
   4.69882798271781
   4.48751281746003
   2.09260761120512
...
 (272 rows)

This example shows how you can use the PREDICT_SVM_REGRESSOR function on the faithful table, using the match_by_pos parameter. In this example, the waiting column was replaced with the constant 40:


=> SELECT PREDICT_SVM_REGRESSOR(40 USING PARAMETERS model_name='mySvmRegModel', match_by_pos='true')
           FROM faithful ORDER BY id;
 PREDICT_SVM_REGRESSOR
--------------------
   1.31778533859324
   1.31778533859324
   1.31778533859324
   1.31778533859324
   1.31778533859324
   1.31778533859324
   1.31778533859324
   1.31778533859324
   1.31778533859324
...
 (272 rows)

5.25 - PREDICT_TENSORFLOW

Applies a TensorFlow model on an input relation, and returns with the result expected for the encoded model type.

Syntax

PREDICT_TENSORFLOW ( input-columns
        USING PARAMETERS model_name = 'model-name' [, num_passthru_cols = 'n-first-columns-to-ignore'] )
OVER( [window-partition-clause] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name: Name of the model (case-insensitive).
num_passthru_cols: Integer that specifies the number of input columns to skip.

Examples

Use PREDICT_TENSORFLOW with the num_passthru_cols parameter to skip the first two input columns:


=> SELECT PREDICT_TENSORFLOW ( pid,label,x1,x2
        USING PARAMETERS model_name='spiral_demo', num_passthru_cols=2 )
OVER(PARTITION BEST) as predicted_class FROM points;

--example output, the skipped columns are displayed as the first columns of the output
  pid  | label |         col0         |         col1
-------+-------+----------------------+----------------------
     0 |     0 |    0.990638732910156 |  0.00936129689216614
     1 |     0 |    0.999036073684692 | 0.000963933940511197
     2 |     1 |   0.0103802494704723 | 0.989619791507721

5.26 - PREDICT_TENSORFLOW_SCALAR

Applies a TensorFlow model on an input relation, and returns with the result expected for the encoded model type. This function supports 1D complex types as input and output.

Applies an imported TensorFlow model on an input relation. This function, unlike PREDICT_TENSORFLOW, accepts one input column of type ROW, where each field corresponds to an input tensor, and returns one output column of type ROW, where each field corresponds to an output tensor.

For details about importing TensorFlow models into Vertica, see TensorFlow integration and directory structure.

Syntax

PREDICT_TENSORFLOW_SCALAR ( inputs
        USING PARAMETERS model_name = 'model-name' )

Arguments

inputs: Input column of type ROW with fields of 1D ARRAYs that represent input tensors. These tensors can represent outputs for various input operations.

Parameters

model_name: Name of the model (case-insensitive).

Examples

This function can simplify the process for making predictions on data with many input features.

For instance, the MNIST handwritten digit classification dataset contains 784 input features for each input row, one feature for each pixel in the images of handwritten digits. The PREDICT_TENSORFLOW function requires that each of these input features are contained in a separate input column. By encapsulating these features into a single ARRAY, the PREDICT_TENSORFLOW_SCALAR function only needs a single input column of type ROW, where the pixel values are the array elements for an input field:

--Each array for the "image" field has 784 elements. 
=> SELECT * FROM mnist_train;
id |                   inputs
---+---------------------------------------------
 1 | {"image":[0, 0, 0,..., 244, 222, 210,...]}
 2 | {"image":[0, 0, 0,..., 185, 84, 223,...]}
 3 | {"image":[0, 0, 0,..., 133, 254, 78,...]}
 ...

In this case, the function output consists of a single opeartion with one tensor. The value of this field is an array of ten elements, which are all zero except for the element whose index is the predicted digit:

=> SELECT id, PREDICT_TENSORFLOW_SCALAR(inputs USING PARAMETERS model_name='tf_mnist_ct') FROM mnist_test;
 id |                   PREDICT_TENSORFLOW_SCALAR                      
----+-------------------------------------------------------------------
  1 | {"prediction:0":["0", "0", "0", "0", "1", "0", "0", "0", "0", "0"]} 
  2 | {"prediction:0":["0", "1", "0", "0", "0", "0", "0", "0", "0", "0"]} 
  3 | {"prediction:0":["0", "0", "0", "0", "0", "0", "0", "1", "0", "0"]} 
...

To view the expected input and output tensors for an imported TensorFlow model, call GET_MODEL_SUMMARY:

=> SELECT GET_MODEL_SUMMARY(USING PARAMETERS model_name='tf_mnist_ct');
                             GET_MODEL_SUMMARY
---------------------------------------------------------------------------

=============
input_tensors
=============
  name |type |dimensions
-------+-----+----------
 image |int32|  [-1,784]  


==============
output_tensors
==============
     name     | type |dimensions
--------------+------+----------
 prediction:0 |int32 | [-1,10]  

(1 row)

5.27 - PREDICT_XGB_CLASSIFIER

Applies an XGBoost classifier model on an input relation.

Applies an XGBoost classifier model on an input relation. PREDICT_XGB_CLASSIFIER returns a VARCHAR data type that specifies one of the following, as determined by how the type parameter is set:

The predicted class (based on probability scores)
Probability of a class for each input instance.

Syntax

PREDICT_XGB_CLASSIFIER ( input-columns
          USING PARAMETERS model_name = 'model-name'
              [, type = 'prediction-type' ]
              [, class = 'user-input-class' ]
              [, match_by_pos = 'match-by-position' ]
              [, probability_normalization = 'prob-normalization' ] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

type

Type of prediction to return, one of the following:

response (default): The class with the highest probability among all possible classes.
probability: Valid only if the class parameter is set, returns for each input instance the probability of the specified class or predicted class.

class

Class to use when the type parameter is set to probability. If you omit this parameter, the function uses the predicted class—the one with the highest probability score. Thus, the predict function returns the probability that the input instance belongs to the specified or predicted class.

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

probability_normalization

The classifier's normalization method, either softmax (multi-class classifier) or logit (binary classifier). If unspecified, the default logit function is used for normalization.

Examples

Use PREDICT_XGB_CLASSIFIER to apply the classifier to the test data:

=> SELECT PREDICT_XGB_CLASSIFIER (Sepal_Length, Sepal_Width, Petal_Length, Petal_Width
    USING PARAMETERS model_name='xgb_iris', probability_normalization='logit') FROM iris1;
PREDICT_XGB_CLASSIFIER
------------------------
setosa
setosa
setosa
.
.
.
versicolor
versicolor
versicolor
.
.
.
virginica
virginica
virginica
.
.
.

(90 rows)

See XGBoost for classification for more examples.

5.28 - PREDICT_XGB_CLASSIFIER_CLASSES

Applies an XGBoost classifier model on an input relation and returns the probabilities of classes:.

Applies an XGBoost classifier model on an input relation and returns the probabilities of classes:

VARCHAR predicted column contains the class label with the highest probability.
Multiple FLOAT columns, where the first probability column contains the probability for the class reported in the predicted column. Other columns contain the probability of each class specified in the classes parameter.
Key columns with the same value and data type as matching input columns specified in parameter key_columns.

All trees contribute to a predicted probability for each response class, and the highest probability class is chosen.

Syntax

PREDICT_XGB_CLASSIFIER_CLASSES ( predictor-columns)
        USING PARAMETERS model_name = 'model-name'
            [, key_columns = 'key-columns']
            [, exclude_columns = 'excluded-columns']
            [, classes = 'classes']
            [, match_by_pos = match-by-position]
            [, probability_normalization = 'prob-normalization' ] )
OVER( [window-partition-clause] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

key_columns

exclude_columns

Comma-separated list of columns from predictor-columns to exclude from processing.

classes

Comma-separated list of class labels in the model. The probability of belonging to each given class is predicted by the classifier. Values are case sensitive.

match_by_pos

Boolean value that specifies how predictor columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the predictor columns list.

probability_normalization

The classifier's normalization method, either softmax (multi-class classifier) or logit (binary classifier). If unspecified, the default logit function is used for normalization.

Examples

After creating an XGBoost classifier model with XGB_CLASSIFIER, you can use PREDICT_XGB_CLASSIFIER_CLASSES to view the probability of each classification. In this example, the XGBoost classifier model "xgb_iris" is used to predict the probability that a given flower belongs to a species of iris:

=> SELECT PREDICT_XGB_CLASSIFIER_CLASSES(Sepal_Length, Sepal_Width, Petal_Length, Petal_Width
    USING PARAMETERS model_name='xgb_iris') OVER (PARTITION BEST) FROM iris1;
  predicted  |    probability
------------+-------------------
 setosa     |   0.9999650465368
 setosa     |   0.9999650465368
 setosa     |   0.9999650465368
 setosa     |   0.9999650465368
 setosa     | 0.999911552783011
 setosa     |   0.9999650465368
 setosa     |   0.9999650465368
 setosa     |   0.9999650465368
 setosa     |   0.9999650465368
 setosa     |   0.9999650465368
 setosa     |   0.9999650465368
 setosa     |   0.9999650465368
 versicolor |  0.99991871763563
 .
 .
 .
(90 rows)

You can also specify additional classes. In this example, PREDICT_XGB_CLASSIFIER_CLASSES makes the same prediction as the previous example, but also returns the probability that a flower belongs to the specified classes "virginica" and "versicolor":

=> SELECT PREDICT_XGB_CLASSIFIER_CLASSES(Sepal_Length, Sepal_Width, Petal_Length, Petal_Width
    USING PARAMETERS model_name='xgb_iris', classes='virginica,versicolor', probability_normalization='logit') OVER (PARTITION BEST) FROM iris1;
 predicted  |    probability    |      virginica       |      versicolor
------------+-------------------+----------------------+----------------------
 setosa     |   0.9999650465368 | 1.16160301545536e-05 | 2.33374330460065e-05
 setosa     |   0.9999650465368 | 1.16160301545536e-05 | 2.33374330460065e-05
 setosa     |   0.9999650465368 | 1.16160301545536e-05 | 2.33374330460065e-05
 .
 .
 .
 versicolor |  0.99991871763563 | 6.45697562080953e-05 |     0.99991871763563
 versicolor | 0.999967282051702 | 1.60052775404199e-05 |    0.999967282051702
 versicolor | 0.999648819964864 |  0.00028366342010669 |    0.999648819964864
 .
 .
 .
 virginica  | 0.999977039257386 |    0.999977039257386 | 1.13305901169304e-05
 virginica  | 0.999977085131063 |    0.999977085131063 | 1.12847163501674e-05
 virginica  | 0.999977039257386 |    0.999977039257386 | 1.13305901169304e-05
(90 rows)

5.29 - PREDICT_XGB_REGRESSOR

Applies an XGBoost regressor model on an input relation.

Applies an XGBoost regressor model on an input relation. PREDICT_XGB_REGRESSOR returns a FLOAT data type that specifies the predicted value by the XGBoost model: a weighted sum of contributions by each tree in the model.

Syntax

PREDICT_XGB_REGRESSOR ( input-columns
        USING PARAMETERS model_name = 'model-name' [, match_by_pos = match-by-position] )

Arguments

input-columns: Comma-separated list of columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name

Name of the model (case-insensitive).

match_by_pos

Boolean value that specifies how input columns are matched to model features:

false (default): Match by name.
true: Match by the position of columns in the input columns list.

Examples

See XGBoost for regression.

5.30 - REVERSE_NORMALIZE

Reverses the normalization transformation on normalized data, thereby de-normalizing the normalized data.

Reverses the normalization transformation on normalized data, thereby de-normalizing the normalized data. If you specify a column that is not in the specified model, REVERSE_NORMALIZE returns that column unchanged.

Syntax

REVERSE_NORMALIZE ( input-columns USING PARAMETERS model_name = 'model-name' );

Arguments

input-columns: The columns to use from the input relation, or asterisk (*) to select all columns.

Parameters

model_name: Name of the model (case-insensitive).

Examples

Use REVERSE_NORMALIZE on the hp and cyl columns in table mtcars, where hp is in normalization model mtcars_normfit, and cyl is not in the normalization model.

=> SELECT REVERSE_NORMALIZE (hp, cyl USING PARAMETERS model_name='mtcars_normfit') FROM mtcars;
hp    | cyl
------+-----
42502 |   8
58067 |   8
26371 |   4
42502 |   8
31182 |   6
32031 |   4
26937 |   4
34861 |   6
34861 |   6
50992 |   8
50992 |   8
49577 |   8
25805 |   4
18447 |   4
29767 |   6
65142 |   8
69387 |   8
14768 |   4
49577 |   8
60897 |   8
94857 |   8
31182 |   6
31182 |   6
30899 |   4
69387 |   8
49577 |   6
18730 |   4
18730 |   4
74764 |   8
17598 |   4
50992 |   8
27503 |   4
(32 rows)

Machine learning functions

Important

1 - Data preparation

1.1 - BALANCE

Behavior type

Syntax

Arguments

Note

Parameters

Privileges

Examples

See also

1.2 - CHI_SQUARED

Tip

Syntax

Arguments

Parameters

Returns

Privileges

Examples

1.3 - CORR_MATRIX

Syntax

Arguments

Returns

Notes

Examples

1.4 - DETECT_OUTLIERS

Note

Behavior type

Syntax

Arguments

Parameters

Privileges

Examples

1.5 - IFOREST

Behavior type

Syntax

Arguments

Parameters

Model Attributes

Privileges

Examples

See also

1.6 - IMPUTE

Behavior type

Syntax

Arguments

Parameters

Privileges

Examples

See also

1.7 - NORMALIZE

Note

Behavior type

Syntax

Arguments

Parameters

Privileges

Examples

See also

1.8 - NORMALIZE_FIT

Note

Behavior type

Syntax

Arguments

Parameters

Model attributes

Privileges

Examples

See also

1.9 - ONE_HOT_ENCODER_FIT

Behavior type

Syntax

Arguments

Parameters

Model attributes

Privileges

Examples

See also

1.10 - PCA