This function accepts one or two input sources, the primary and secondary input. Behavior differs depending on the number of inputs passed in. The primary input is a SERIES_SPEC which references univariate series instances. These are the series whose coefficients are being estimated by the function. The payload content for that input source is always REAL. The optional secondary input is either an ART_SPEC or SERIES_SPEC which references instances of binary strings that originated from an ARTMODEL layer constructed by a previous TD_ARIMAESTIMATE with FIT_PERCENTAGE(100) or TD_ARIMAVALIDATE invocation. These binary strings are the previously-generated models that are applied by the function. The payload content for the second input is MULTIVAR_ANY.
When only the primary input is passed in, the function estimates the coefficients associated with the ARIMA model. When both a primary and secondary inputs are passed in, the function applies the model referenced by the secondary input to the primary series. The end result is an ART output that can be passed to the TD_ARIMAFORECAST function for forecasting processing.