9. MULTIVARIATE REGRESSION FORECASTING › 9.3 Standard Output › 9.3.1 Update to the Forecast File

9.3.1 Update to the Forecast File

You can use the Save forecast option on the Multivariate
Regression Forecasting control screen shown in Figure 9-10 to
save forecasted values of the dependent variable in the
forecast file that corresponds to the file containing the
dependent variable.

Typically, as a user of Multivariate Regression Forecasting,
you can develop a model using the three steps outlined in
Section 9.2.  This is almost always an iterative process, for
the first few choices of model parameters may not be the best
choices.  Formulate the model and then review the results of
that model using the three tests described in Section 9.2.3.
Should the tests fail, make corrections and adjustments to
the proposed model and rerun the forecast.

Finally, you develop a model that satisfies your
requirements.  At this point, you can use the model to
generate future values for the dependent variable.  To do
this, future values of ALL the independent variables must
exist.  By rerunning the same model with the Save forecast
option mentioned above, the forecasting process generates the
future values of the dependent variable and stores them in
the appropriate file.

In the forecast file, future values are stored in a special
way that is used by other functions of the CA MICS Capacity
Planner to produce comparative graphs and charts.  The
FILE_NO variable is used to distinguish between the various
types of observations within forecast files.  The
Multivariate Regression, Univariate Model Forecasting, and
Business Element Forecasting facilities store a future value
with a FILE_NO of 11, and a confidence range with a FILE_NO
value of 12.  In addition, each observation of a forecast
contains the variables FORYEAR and FORMNDY, FORWEEK or
FORMONTH, which contain the date for the most recent
historical data used in making the forecast.  These concepts
are discussed in more detail in Section 3.2.