1.0 - 8.00 - VARMAX Output - Teradata Vantage

Teradata® Vantage Machine Learning Engine Analytic Function Reference

Product
Teradata Vantage
Release Number
1.0
8.00
Release Date
May 2019
Content Type
Programming Reference
Publication ID
B700-4003-098K
Language
English (United States)

The VARMAX function outputs a model for each partition in the input table.

Output Schema

Column Data Type Description
partition_column Same as in input table Identifier of partition_column from input table.
coef VARCHAR Value from following table.
coef_value VARCHAR Value explanation from following table.
stepahead Integer [Appears only if StepAhead argument has positive value.] Identifier of future period. For example, for StepAhead (3), this column contains 1, 2, and 3 (in separate rows) to indicate the future period for which a predicted value appears in the next column.
predict_columnname Double [Appears only if StepAhead argument has positive value.] [Column appears once for each specified response_column.] Predicted future values of response variable.

Model Coefficients

coef coef_value
coef Vector [p, d, q, sp, sd, sq, b].
ar_params Matrixes Φi, as vector of p matrixes, each of which is an n *n matrix. p is from coef vector and n is number of response variables specified in ResponseColumns argument.
ma_params Matrixes Φi, as vector of q matrixes, each of which is an n *n matrix. q is from coef vector and n is number of response variables specified in ResponseColumns argument.
exogenous_params Matrixes B i , shown as a vector of b matrixes, each of which is an n *m matrix. b is from coef vector, n is number of response variables specified in ResponseColumns argument, and m is number of exogenous variables specified in ExogenousColumns argument.
seasonal_ar_params Matrix Φsi for seasonal parameters.
seasonal_ma_params Matrix Φsi for seasonal parameters.
mean_param [Appears only with IncludeMean ('true').] Mean vector of response series.
period Cycle period for seasonal models (0 for nonseasonal models).
lag Lag value specified in function call.
sigma The variance matrix.
aic Akaike information criterion.
bic Bayesian information criterion.
iterations Number of iterations performed.
converged Whether algorithm converged.