9.3.2.2 Residual Analysis Report

9. MULTIVARIATE REGRESSION FORECASTING › 9.3 Standard Output › 9.3.2 Standard Reports › 9.3.2.2 Residual Analysis Report

9.3.2.2 Residual Analysis Report


The Residual Analysis Report shows how the multivariate
regression model fits the historical data series.  Figure 9-2
shows a sample Residual Analysis Report.

CA MICS Capacity Planner MULTIVARIATE REGRESSION FORECAST AND RESIDUAL ANALYSIS MULTIVARIATE REGRESSION BASED FORECAST OF: TOTCPUTM MODEL BASED ON INDEPENDENT ELEMENTS : TSOCPUTM BATCPUTM IMSCPUTM CICCPUTM 95% INDEPENDENT-ELEMENT PREDICTED RESIDUAL -------------PLOT OF RESIDUALS------------- CONFIDENCE DATE TOTCPUTM NAME VALUE TOTCPUTM TOTCPUTM -18723 0.0 ퟎ LIMITS ------- ---------- -------- --------- ---------- ---------- ------------------------------------------- ---------- 04JAN97 95:34:40.7 TSOCPUTM 47321 91:47:15.0 3:47:25.7 | 0ﯯ뻻ﯯ | 1:26:03.7 BATCPUTM 37461 | 0 | IMSCPUTM 118575 | 0 | CICCPUTM 60716 | 0 | | 0 | 11JAN97 91:11:54.9 TSOCPUTM 47779 90:36:08.0 0:35:46.9 | 0 | 1:20:01.3 BATCPUTM 37663 | 0 | IMSCPUTM 115851 | 0 | CICCPUTM 59529 | 0 | | 0 | 18JAN97 93:57:10.4 TSOCPUTM 49732 92:14:06.2 1:43:04.2 | 0ﯯ | 1:04:26.9 BATCPUTM 36463 | 0 | IMSCPUTM 119521 | 0 | CICCPUTM 59195 | 0 | | 0 | 25JAN97 92:25:59.9 TSOCPUTM 52382 93:46:39.6 -1:20:39.6 | -----0 | 1:02:14.1 BATCPUTM 35767 | 0 | IMSCPUTM 121372 | 0 | CICCPUTM 59275 | 0 | | 0 | 01FEB97 94:15:30.1 TSOCPUTM 52535 94:58:53.1 -0:43:22.9 | --0 | 1:03:10.5 BATCPUTM 36859 | 0 | IMSCPUTM 122859 | 0 | CICCPUTM 59992 | 0 | | 0 | 08FEB97 96:10:44.4 TSOCPUTM 54022 97:40:15.2 -1:29:30.8 | -----0 | 1:08:11.3 BATCPUTM 36984 | 0 | IMSCPUTM 126560 | 0 | CICCPUTM 61944 | 0 |


 Figure 9-2.  Residual Analysis Report

The Residual Analysis Report includes the following
information:

MULTIVARIATE
REGRESSION
BASED         The dependent data element name that you
FORECAST OF:  specify on the entry screen.

MODEL         The independent elements that you use to
BASED ON      develop the model.  This list may differ from
INDEPENDENT   the list of independent elements specified in
ELEMENTS:     the Model Analysis Report shown in Figure 9-1.
              This difference results from deleting proposed
              independent elements based on the minimum
              r-squared improvement criterion.

The columns of the report are discussed below:

DATE:         The date ending value for the week or month
              represented by the observation.

Independent   The actual historical observations for the data
Element       element being modeled.  Note that if the last
Name          two characters of the data element name are TM,
(TOTCPUTM):   the program assumes that the value is in
              seconds and uses a SAS TIME10.1 format for the
              report.

Two columns appear under the heading INDEPENDENT-ELEMENT for
each independent element used in the model.

NAME:         The name of the independent element.

VALUE:        The value of the independent element for the
              interval.

The remaining columns of the report are described as follows:

PREDICTED     The predicted value for the dependent variable
(TOTCPUTM):   for the interval using the regression equation
              fit to the historical data.  Note that if the
              last two characters of the data element name
              are TM, the program assumes that the value is
              in seconds and uses a SAS TIME10.1 format for
              the report.

RESIDUAL      The error term in the regression equation.  The
(TOTCPUTM):   residual is equal to the difference between the
              actual historical observation and the predicted
              value.  The residual value for a point deleted
              from the modeling process is of no concern
              since the model was not developed to account
              for the observation.  Note that if the last two
              characters of the data element name are TM,
              then the program assumes that the value is in
              seconds and uses a SAS TIME10.1 format for the
              report.

PLOT OF       The residual values plotted over the range from
RESIDUALS:    minus to plus the maximum residual value.

CONFIDENCE    The confidence limits for the forecast.  The
LIMITS:       confidence limits bound the potential error
              that may exist in the forecasts produced by the
              model.  The confidence limits are a function of
              the independent element estimates provided,
              rather than monotonically increased as was the
              case with models developed by the Univariate
              Model Forecasting.  If the last two characters
              of the data element name are TM, the program
              assumes that the value is in seconds and uses a
              SAS TIME10.1 format for the report.