Data Mining Techniques for Segmentation - Data Mining Techniques in CRM: Inside Customer Segmentation - page 117

Figure 3.26 A decision tree model in IBM SPSS Modeler (Ruleset format).
Handling of Predictors by Decision Tree Models
Decision tree models can handle both categorical and numeric continuous
predictors.
The categories of each categorical predictor are assessed with respect
to the output and the homogeneous groups are merged (regrouped) before
evaluation of the specific categorical predictor for splitting. As an example,
let us consider a hypothetical churn prediction model which assesses the