Previous Topic: 9.3.2 Standard ReportsNext Topic: 9.3.2.2 Residual Analysis Report


9.3.2.1 Model Analysis Report


Figure 9-1 presents a sample Model Analysis Report for
Multivariate Regression Forecasting.  The Model Analysis
Report provides a summary of the regression analysis that was
conducted for the specified multivariate regression model.

CA MICS Capacity Planner ANALYSIS OF MULTIVARIATE REGRESSION MODELS MODEL OF: TOTCPUTM BASED ON: TSOCPUTM IMSCPUTM CICCPUTM --------INDEPENDENT--ELEMENTS-------- R**2 F P INTERCEPT NAME COEFFICIENT F P -------- ------- ----- --------- -------- ----------- ------- ----- 0.97 1127.97 .0001 10:57:58.3 TSOCPUTM 1.77499 71.77 .0001 IMSCPUTM 1.14392 77.90 .0001 CICCPUTM 1.13065 31.88 .0001


 Figure 9-1.  Model Analysis Report

The Model Analysis Report contains the following information:

MODEL OF:     The dependent data element that you wish to
              forecast.

BASED ON:     The independent data elements on which the
              analysis is based.

The first four columns of the report are discussed below.

R**2:         The r-squared value for the model.  Although
              the SAS REG procedure attempts to use every
              one of the independent elements suggested by
              the analyst, the procedure excludes any term
              that does not result in a minimum improvement
              in the r-squared value as specified on the
              Multivariate Regression Forecasting screen. We
              recommend an r-squared value of 0.70 and above
              for the acceptance of models produced with
              Multivariate Modeling Forecasting.

F:            The F statistic for the model.  As discussed
              in Section 9.2.3, the F statistic is calculated
              from the ratio of sums of squares of the model.
              Larger F values indicate a more reliable model.
              The F statistics may be unreliable for models
              developed with a small number of points (that
              is, less than 30).  An F value is provided for
              each step made by the regression procedure.

P:            The probability that the model proposed is
              significant (that is, more reliable than
              forecasting the average value of the historical
              observations).  Although most statistics texts
              recommend testing the null hypothesis to a
              0.001 level, we recommend 0.01 for models based
              on computer measurement data.  A p value is
              provided for each step made by the regression
              procedure.

INTERCEPT:    The b value for the regression equation.

The next four columns are produced for each independent
element that you specify.  These columns contain the
following information:

NAME:         The name of the independent element being
              analyzed.

COEFF:        The coefficient for the independent element in
              the regression equation.  The coefficient is
              the M value mentioned in Sections 9.2.1 and
              9.4.  Negative coefficients indicate that you
              can expect resource consumption to decrease
              with increases in the independent element
              value.  This situation is usually not normal in
              resource consumption models.

F:            The F statistics calculated for the business
              element.  This F value indicates the
              significance of the independent element in the
              model that was developed.  Once again, larger
              values indicate a more reliable model and the
              value may be misleading for models based on
              small samples.

P:            The probability that the specified business
              element makes a significant contribution to the
              model developed.  We recommend a 0.01
              acceptance value.

This report is used to evaluate whether the model that is
produced adequately models an application.  Although the
values of r-squared, F, and p indicate the statistical
significance of the proposed model, remember that a proper
evaluation includes all three tests enumerated in Section
9.2.3: statistical interpretation, visual inspection, and
the common sense test.