Advanced Modeling and Optimization

Abstract for Paper 4 of Volume 4, Number 1, 2002, pp. 39-56

Classification for three-group of credit cardholders’ behavior via a multiple criteria approach.


Yi Peng
College of Information Science and Technology,
University of Nebraska at Omaha
Omaha, NE 68182, USA
Yong Shi
College of Information Science and Technology,
University of Nebraska at Omaha
Omaha, NE 68182, USA
Weixuan Xu
Institute of Policy and Management,
Chinese Academy of Sciences, Beijing 100080, China
Abstract

Classification is a main form of data mining that can be used to derive models describing important data classes. The problem of classification has strong impact on many business areas, such as marketing analysis, credit card portofolio management, and insurance actuarial research. This paper proposes a multiple criteria approach to classify three groups (e.g. bad, normal, or good) credit cardholder behaviors using the First Data Resource's credit score. Based on the preliminary research findings of Shi et al (2001) about two-group (bad or good) credit cardholder behaviors, the three-group multi-criteria model is represented and tested using real-life data. The SAS linear programming package is utilized to facilitate the computation of the model.