Predictive Modeling : Scorecard Development in Credit Risk Management - a technical paper using decision-tree model

CHAID (chi-square automatic interaction detection) offers a method of efficiently searching for relations between the predictor variables and a categorical outcome measure, and reporting in a tree-diagram format the category combinations that make the largest difference in the outcome percentages. Thus CHAID is an effective tool for identifying segments that maximize the positive response (purchase, renewal) you are seeking. This case study applies CHAID procedure in SPSS 15.0 to build a tree-based decision model to develop the score card system for a bank. The SPSS 15.0 syntax was attached in the Appendix.

 Report