General Classification/Regression Tree Models Syntax - Overview

Syntax for C&RT analyses. The syntax for specifying the C&RT analysis is similar in many ways to the syntax for specifying General Linear Models. In particular, in order to analyze ANOVA/ANCOVA-like coded designs, you can use the same command syntax as you would use for specifying a linear model. Thus, by using the command syntax it is very easy to move between different methods of analyzing data (e.g., General Linear Models, General Discriminant Analysis models, Generalized Linear/Nonlinear Models, General CHAID Models, etc.) to evaluate the predictive power of the same variables (or design of the predictor effects) with different methods. The differences in the syntax pertain to the specifications of estimation options, which will be different for different techniques. Those keywords specific to C&RT are listed in Syntax Editor - Keywords tab. Refer to the GLM Syntax for details concerning how to specify designs.

Using  the Quick specs dialog to write the syntax. In practically all cases, the most efficient way to write C&RT syntax is to use the Quick specs dialog box, and then click the Syntax editor button in that dialog box to "translate" the selected specifications to the command syntax. This can further be edited and updated as and when required.