- primary_art_specification
- An analytical result table (ART) that was populated by TD_ARIMAXESTIMATE that produced the ARTMODEL layer containing the ARIMAX model.
- secondary_series_specification
- The runtime series for the future exogenous variables (time series) that is passed to the TD_ARIMAXFORECAST function. Only supports real numbers.
- FUNC_PARAMS
-
Name Data Type Description FORECAST_PERIODS INTEGER Number of steps (h) to forecast ahead. If using EXOG_IMPUTE (NONE), then h must be equal to number of observations in the second input, otherwise an error occurs"Number of entries for exogenous variables should match with FORECAST_PERIODS."PREDICTION_INTERVALS Enumerated, String [Optional] The confidence level for the prediction, such that 85 means 85% confidence. Options are: - NONE
- 85
- 90
- BOTH
Default is BOTH.
EXOG_IMPUTE Enumerated, String [Optional] Data to use as future values of exogenous variables during the endogenous variable forecasting. Options are: - NONE: Use actual values of FORECAST_PERIODS in second input. This allows pre-processing to get future values and requires number of entries to equal to number of FORECAST_PERIODS.
- MEAN: Use mean of second input for the endogenous variable forecasting.
- MEDIAN: Use median of second input for the endogenous variable forecasting.
Default is NONE.
- INPUT_FMT
- [Optional] The INPUT_MODE parameter has the following options:
- ONE2ONE: Both the primary and secondary series specifications contain a series name which identifies the two series in the function.
- MANY2ONE: The MANY specification is the primary series declaration. The secondary series specification contains a series name that identifies the single secondary series.
- MATCH: Both series are defined by their respective SERIES_SPEC(SERIES_ID()) declarations.
- OUTPUT_FMT
- Output format is INDEX_STYLE, and the only index style supported is NUMERICAL_SEQUENCE.