Time Series Model Package
TSMFOR Object
The TSMFOR object collects forecast series from a TSM object, TSMObj, and stores them in an output table. The table schema that is used for storing the set of forecast series variables is compatible with the schema that the HPFENGINE procedure uses for the data set that is specified in the TSMFOR= option. Alternatively, the TSMFOR object can be configured to collect from a TSM object the smoothed (imputed) values for the dependent series and their associated standard errors and confidence limits.
Table 15 shows the contents of the TSMFOR object.
Table 15: Contents of the TSMFOR Object
| Column | Type | Description |
|---|---|---|
| _NAME_ | String | Name of the dependent variable |
| _MODEL_ | String | (Optional) Model name |
| _TIMEID_ | Numeric | Uniform time ID values for series |
| ACTUAL | Numeric | Accumulated values of dependent variable |
| ERROR | Numeric | Residuals |
| LOWER | Numeric | Lower confidence limit |
| PREDICT | Numeric | Forecasts of the dependent variable |
| STD | Numeric | Prediction standard error |
| UPPER | Numeric | Upper confidence limit |
Table 16 summarizes the methods that are associated with the TSMFOR object.
Table 16: Methods of the TSMFOR Object
Figure 11 illustrates the data flow through the TSMFOR object.
Figure 11: TSMFOR Object Data Flow
