Chapter Contents |
Previous |
Next |

The ARIMA Procedure |

The OUTMODEL= option in the ESTIMATE statement writes an output data set that enables you to reconstruct the model. The OUTMODEL= data set contains much the same information as the OUTEST= data set but in a transposed form that may be more useful for some purposes. In addition, the OUTMODEL= data set includes the differencing operators.

The OUTMODEL data set contains the following:

- the BY variables
- _NAME_, a character variable containing the name of the response or input variable for the observation.
- _TYPE_, a character variable that contains the estimation method that was employed. The value of _TYPE_ can be CLS, ULS, or ML.
- [_STATUS_] This variable describes the convergence status of the model. A value of 0_CONVERGED indicates that the model converged.
- _PARM_, a character variable containing the name of the parameter given by the observation. _PARM_ takes on the values ERRORVAR, MU, AR, MA, NUM, DEN, and DIF.
- _VALUE_, a numeric variable containing the value of the estimate defined by the _PARM_ variable.
- _STD_, a numeric variable containing the standard error of the estimate.
- _FACTOR_, a numeric variable indicating the number of the factor to which the parameter belongs.
- _LAG_, a numeric variable containing the number of the term within the factor containing the parameter.
- _SHIFT_, a numeric variable containing the shift value for the input variable associated with the current parameter.

The values of _FACTOR_ and _LAG_ identify which particular MA, AR, NUM, or DEN parameter estimate is given by the _VALUE_ variable. The _NAME_ variable contains the response variable name for the MU, AR, or MA parameters. Otherwise, _NAME_ contains the input variable name associated with NUM or DEN parameter estimates. The _NAME_ variable contains the appropriate variable name associated with the current DIF observation as well. The _VALUE_ variable is 1 for all DIF observations, and the _LAG_ variable indicates the degree of differencing employed.

The observations contained in the OUTMODEL= data set are identified by the _PARM_ variable. A description of the values of the _PARM_ variable follows:

- NUMRESID
- _VALUE_ contains the number of residuals.
- NPARMS
- _VALUE_ contains the number of parameters in the model.
- NDIFS
- _VALUE_ contains the sum of the differencing lags employed
for the response variable.
- ERRORVAR
- _VALUE_ contains the estimate of the innovation variance.
- MU
- _VALUE_ contains the estimate of the mean term.
- AR
- _VALUE_ contains the estimate of the autoregressive parameter
indexed by the _FACTOR_ and _LAG_ variable values.
- MA
- _VALUE_ contains the estimate of a moving average parameter
indexed by the _FACTOR_ and _LAG_ variable values.
- NUM
- _VALUE_ contains the estimate of the parameter in the numerator factor
of the transfer function of the input variable
indexed by the _FACTOR_, _LAG_, and _SHIFT_ variable values.
- DEN
- _VALUE_ contains the estimate of the parameter in the denominator factor
of the transfer function of the input variable
indexed by the _FACTOR_, _LAG_, and _SHIFT_ variable values.
- DIF
- _VALUE_ contains the difference operator defined by the difference lag given by the value in the _LAG_ variable.

Chapter Contents |
Previous |
Next |
Top |

Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.