Model Parameters and Forecasts (Temporal Causal Model Forecasting)
- Load from model file
- Forecasts are produced by using the model parameters from the model system file, and the data from the active dataset, without reestimating the model parameters. Goodness of fit measures that are displayed in output and used to select best-fitting models are taken from the model system file. The fit measures then reflect the data that was used when each model was developed (or last updated). This option is appropriate for generating forecasts and output from the data that was used to build the model system.
- Reestimate from data
-
Model parameters are reestimated by using the data in the active dataset. Reestimation of model parameters does not affect which inputs are included in the model for each target. This option is appropriate when you have new data beyond the original estimation period and you want to generate forecasts or other output with the updated data.
- All observations
- Specifies that the estimation period starts at the time of the earliest observation and ends at the time of the latest observation across all series.
- By start and end times
- You can specify both the start and end of the estimation period or you can
specify just the start or just the end. If you omit the start or the end of the
estimation period, the default value is used.
- If the observations are defined by a date/time field, then enter values for start and end in the same format that is used for the date/time field.
- For observations that are defined by cyclic periods, specify a value for each of the cyclic periods fields. Each field is displayed in a separate column.
- If there is a date specification in effect for the active dataset, then you must specify a value for each component (such as Month) of the date specification. Each component is displayed in a separate column.
- When the observations are defined by record order, the start and end of the estimation period are defined by the row number (as displayed in the Data Editor) of the relevant case.
- By latest or earliest time intervals
-
Defines the estimation period as a specified number of time intervals that start at the earliest time interval or end at the latest time interval in the data, with an optional offset. In this context, the time interval refers to the time interval of the analysis. For example, assume that the observations are monthly but the time interval of the analysis is quarters. Specifying Latest and a value of 24 for the Number of time intervals means the latest 24 quarters.
Optionally, you can exclude a specified number of time intervals. For example, specifying the latest 24 time intervals and 1 for the number to exclude means that the estimation period consists of the 24 intervals that precede the last one.
- Extend records into the future
- Sets the number of time intervals to forecast beyond the end of the estimation period. The time interval in this case is the time interval of the analysis. When forecasts are requested, autoregressive models are automatically built for any input series that are not also targets. These models are then used to generate values for those input series in the forecast period to obtain the forecasts for the targets of those inputs.