Temporal Causal Model Forecasting

The Temporal Causal Model Forecasting procedure loads a model system file that was created by the Temporal Causal Modeling procedure and applies the models to the active dataset. You can use this procedure to obtain forecasts for series for which more current data is available, without rebuilding your models. You can also generate any of the output that is available with the Temporal Causal Modeling procedure.

Assumptions

  • The structure of the data in the active dataset, either column-based or multidimensional, must be the same structure that was used when the model system was built. For multidimensional data, the dimension fields must be the same as used to build the model system. Also, the dimension values that were used to build the model system must exist in the active dataset.
  • Models are applied to fields in the active dataset with the same names as the fields specified in the model system.
  • The field, or fields, that defined the observations when the model system was built must exist in the active dataset. The time interval between observations is assumed to be the same as when the models were built. If the observations were defined by a date specification, then the same date specification must exist in the active dataset. Date specifications are created from the Define Dates dialog or the DATE command.
  • The time interval of the analysis and any settings for aggregation, distribution, and missing values are the same as when the models were built.

This procedure pastes TCM APPLY command syntax.