A Multilayered Approach to Classifying Customer Responsiveness and Credit Risk

TL;DR

This paper compares multiple classifiers across response, risk, and combined models to improve credit card campaign targeting and default prediction, highlighting the effectiveness of Extra Trees and Random Forest classifiers.

Contribution

It introduces a multilayered classification approach optimizing different models for response and risk prediction in credit card campaigns.

Findings

Extra Trees achieved 79.1% recall in response model

Random Forest achieved 84.1% specificity in risk model

Random Forest achieved 83.2% accuracy in response-risk model

Abstract

This study evaluates the performance of various classifiers in three distinct models: response, risk, and response-risk, concerning credit card mail campaigns and default prediction. In the response model, the Extra Trees classifier demonstrates the highest recall level (79.1%), emphasizing its effectiveness in identifying potential responders to targeted credit card offers. Conversely, in the risk model, the Random Forest classifier exhibits remarkable specificity of 84.1%, crucial for identifying customers least likely to default. Furthermore, in the multi-class response-risk model, the Random Forest classifier achieves the highest accuracy (83.2%), indicating its efficacy in discerning both potential responders to credit card mail campaign and low-risk credit card users. In this study, we optimized various performance metrics to solve a specific credit risk and mail responsiveness business problem.