A set of data mining models to classify credit cardholder behavior

  • Authors:

  • Gang Kou;Yi Peng;Yong Shi;Weixuan Xu

  • Affiliations:

  • College of Information Science and Technology, University of Nebraska at Omaha, Omaha, NE;College of Information Science and Technology, University of Nebraska at Omaha, Omaha, NE;College of Information Science and Technology, University of Nebraska at Omaha, Omaha, NE;Institute of Policy and Management, Chinese Academy of Sciences, Beijing, China

  • Venue:
  • ICCS'03 Proceedings of the 2003 international conference on Computational science: PartII
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, we present a set of classification models by using multiple criteria linear programming (MCLP) to discover the various behaviors of credit cardholders. In credit card portfolio management, predicting the cardholder's spending behavior is a key to reduce the risk of bankruptcy. Given a set of predicting variables (attributes) that describes all possible aspects of credit cardholders, we first present a set of general classification models that can theoretically handle any size of multiple-group cardholders' behavior problems. Then, we implement the algorithm of the classification models by using SAS and Linux platforms. Finally, we test the models on a special case where the cardholders' behaviors are predefined as five classes: (i) bankrupt charge-off; (ii) non-bankrupt charge-off; (iii) delinquent; (iv) current and (v) outstanding on a real-life credit card data warehouse. As a part of the performance analysis, a data testing comparison between the MCLP and induction decision tree approaches is demonstrated. These findings suggest that the MCLP-data mining techniques have a great potential in discovering knowledge patterns from a large-scale real-life database or data warehouse.