PLS Syntax - Example 3: Multivariate Regression with SIMPLS

This example illustrates the specification for a SIMPLS (see Introductory Overview) multiple regression problem with two dependent variables and categorical predictors. In this example, we also assume that we are interested in first 10 components of PLS.

Note: brief descriptive comments are enclosed in curly brackets. You can run this example with the example data file Exp.sta.

PLS;

{ Dependent variable list: }

   DEPENDENT = STRESS_R CORRECT3;

{ Specification of grouping variables (factors); note that

  optional grouping codes (values) are specified to identify

  the (selected) groups in the categorical predictors. }

   GROUPS = GENDER ("MALE" "FEMALE")

            TIME   (1 2 3);

{ Specification of continuous predictor variables (covariates) }

   COVARIATE = CORRECT1 CORRECT2;

{ Here the bar operator is used to construct the full factorial

  design for the categorical predictor variables; the bar operator

  will evaluate to all main effects and interactions

  between the grouping variables. }

   DESIGN = CORRECT1+CORRECT2+GENDER|TIME;

{ The maximum number of components to be extracted is set

   to 10. }

   METHOD = SIMPLS;

For more examples, see PLS Syntax - Examples.