SAVE Subcommand (TSMODEL command)

The SAVE subcommand is used to save new variables representing predicted values, residuals, and confidence intervals to the active dataset. By default, no new variables are saved to the active dataset.

  • Specify one or more keywords, each keyword followed by an optional rootname to be used as the prefix for new variable names. Enclose the rootname in parentheses. Each keyword gives rise to one new variable for each dependent variable.
  • The rootname, if specified, must conform to the rules for valid variable names.
  • If no rootname is specified, TSMODEL uses a default name.
  • The full variable name is the concatenation of the rootname, the name of the associated dependent variable, and a model identifier. The variable name is extended if necessary to avoid variable naming conflicts.
  • The PREDICT command controls whether new cases are added to the file when new variables are saved. New cases are added if PREDICT specifies a forecast period that extends beyond the length of the dependent-variable series.
  • SAVE is ignored with a warning if temporary transformations are in effect.

Example

TSMODEL
  /SAVE PREDICTED(Pred) NRESIDUAL
  /MODEL DEPENDENT=sku1 TO sku10.
  • Two new variables are created for each of the dependent variables sku1 thru sku10: one variable contains the model predictions, and the other variable contains the noise residuals.
  • The rootname (prefix) for the variables containing model predictions is Pred.
  • The rootname for the variables containing noise residuals is NResidual (the default).

PREDICTED(rootname). Model fit and forecast values. The default rootname is Predicted.

LCL(rootname). Lower confidence limits. The default rootname is LCL.

UCL(rootname). Upper confidence limits. The default rootname is UCL.

NRESIDUAL(rootname). Noise residuals. The default rootname is NResidual. When transformations of the dependent variable are performed (for example, natural log), these residuals are the residuals for the transformed series.