The Art of Credit Scorecard Development: Part 3— Credit ScoreCard Development — Using Toad Package

This is the final part of this part 3 series of credit scorecard development. Part 1 focused on understanding the business context and preparing data. Part 2 entailed all preliminary techniques preceding scorecard creation using Toad Package. Finally, Part 3, will finalize with model and scorecard development.

Combined Code Notebook link : Github

Objective

We concluded Part 2 by identifying the final features for model development. In this article, we will use these features to build a model, which will ultimately be used to develop a scorecard.

Stepwise Regression

Once we have the final features from Part 2, we will use stepwise regression (forward, backward, and bidirectional) for further feature selection. This statistical technique employs tests like T-test and F-test to identify the most significant features that improve model performance. Common selection criteria include AIC, BIC, and Adjusted R-squared.

• Forward Stepwise: Starts with no predictors adds variables one by one based on their correlation with the dependent variable, and stops when no significant improvement is achieved.
• Backward Stepwise…