The MVPDIAGNOSE Procedure

SCOREMATRIX Statement

  • SCOREMATRIX < / options>;

The SCOREMATRIX statement produces a matrix of score plots, each of which is a scatter plot of scores for a pair of principal components. You can use the SCOREPLOT statement to display a single score plot in a graph by itself.

Table 11.3 summarizes the options available in the SCOREMATRIX statement.

Table 11.3: SCOREMATRIX Statement Options

Option

Description

ALPHA=

Specifies the $\alpha $ value for prediction ellipses

ELLIPSE

Requests prediction ellipses to be overlaid on score plots

GROUP=

Specifies a variable for grouping points in score plots

LABELS=

Specifies whether points in the score plots are labeled

NCOMP=

Specifies the number of principal components whose scores are plotted

ODSFOOTNOTE=

Adds a footnote to the score matrix

ODSFOOTNOTE2=

Adds a secondary footnote to the score matrix

ODSTITLE=

Specifies a title for the score matrix

ODSTITLE2=

Specifies a secondary title for the score matrix


You can specify the following options in the SCOREMATRIX statement. The section Common Plot Statement Options describes additional options that are available in all plot statements.

ALPHA=$\alpha $

specifies the $\alpha $ value for prediction ellipses that are overlaid on the score plots. The probability that a new observation falls outside the ellipse is $\alpha $. The default is 0.05. If you specify the ALPHA= option, you do not need to specify the ELLIPSE option.

ELLIPSE

requests that prediction ellipses be overlaid on the score plots. The probability that a new observation falls outside the ellipse is specified by the ALPHA= option.

GROUP=variable

specifies a variable in the input data set that is used to group the points in the score plots. Points that have different GROUP= values are plotted using different markers or colors (or both) to distinguish the groups.

LABELS=ON | OFF

specifies whether points in the score plots are labeled. Points are labeled with the values of the first variable listed in the ID statement, or the observation number if no ID statement is specified. The default is LABELS=OFF.

NCOMP=n

specifies the number of principal components whose scores are plotted in the matrix. The principal components that are plotted are always 1 through n. By default, the matrix contains score plots for all principal components.