The MVPDIAGNOSE Procedure

PROC MVPDIAGNOSE Statement

PROC MVPDIAGNOSE <options> ;

The PROC MVPDIAGNOSE statement invokes the MVPDIAGNOSE procedure and specifies input data sets. You can specify the following options:

DATA=SAS-data-set

specifies an input SAS data set that contains process measurement data for a Phase II analysis. If you specify a DATA= data set, you must also specify a LOADINGS= data set. You cannot specify both the HISTORY= option and the DATA= option. See the section DATA= Data Set for details about DATA= data sets.

HISTORY=SAS-data-set

specifies an input SAS data set that contains process variable data that are augmented with principal component scores, multivariate summary statistics, and other calculated values. Usually you create a HISTORY= data set by using the OUT= option in the PROC MVPMODEL statement or the OUTHISTORY= option in the PROC MVPMONITOR statement. You cannot specify both the DATA= option and the HISTORY= option. See the section HISTORY= Data Set for details about HISTORY= data sets.

LOADINGS=SAS-data-set

specifies an input SAS data set that contains eigenvalues, principal component loadings, and process variable means and standard deviations that are used to compute principal component scores and multivariate summary statistics for a Phase II analysis. Usually you create a LOADINGS= data set by using the OUTLOADINGS= option in the PROC MVPMODEL statement. See the section LOADINGS= Data Set for details about LOADINGS= data sets.

MISSING=AVG | NONE

specifies how observations that have missing process variable values in the DATA= data set are to be handled. The option MISSING=AVG specifies that missing values for a given variable be replaced by the average of the nonmissing values for that variable. The default is MISSING=NONE, which excludes from the analysis any observation that has missing values for any of the process variables.

PREFIX=name

specifies the prefix that is used to identify variables that contain principal component scores in the HISTORY= data set. For example, if you specify PREFIX=ABC, PROC MVPDIAGNOSE attempts to read the score variables ABC1, ABC2, ABC3, and so on. The default prefix is Prin, which is also the default score variable prefix for data sets created by using the OUT= option in the PROC MVPMODEL statement. If you are using an OUT= data set from PROC MVPMODEL in the HISTORY= data set, the PREFIX= values must match. That is, the PREFIX= value that is specified in the PROC MVPDIAGNOSE statement must match the PREFIX= value in the data set that is specified in the OUT= option in the PROC MVPMODEL statement.

Note: The number of characters in the prefix plus the number of digits that are required to enumerate the principal components must not exceed the current name length defined by the VALIDVARNAME= system option.

RPREFIX=name

specifies the prefix that is used to identify variables that contain residuals in the HISTORY= data set. A residual variable name is formed by appending a process variable name to the prefix. The default prefix is R_, which is also the default residual variable prefix for data sets created by using the OUT= option in the PROC MVPMODEL statement. If you are using a data set produced with the OUT= option in the PROC MVPMODEL statement as a HISTORY= data set, the RPREFIX= value must match the RPREFIX= value specified when the OUT= data set was created by PROC MVPMODEL.

If the combined length of the residual prefix and a process variable name exceeds the maximum name length defined by the VALIDVARNAME= system option, characters are removed from the middle of the process variable name before it is appended to the residual prefix. For example, if you specify RPREFIX=Residual_ (nine characters), the maximum variable name length is 32, and there is a process variable named PrimaryThermometerReading (25 characters), then two characters are dropped from the middle of the process variable name. The resulting residual variable name is Residual_PrimaryThermeterReading.