A Data Mining Approach to Classify Credit Cardholders' Behavior

Credit cardholders' behavior analysis is an important issue to be studied. Multi-criteria linear programming (MCLP) classification method has shown its advantage in fast speed and balanced classification accuracy on this problem. However, dimension reduction is necessary before some classification methods implementing, not only for faster classification speed but also for commercial knowledge discovering. In this paper, a data mining approach based on the combination of MCLP and principal component analysis (PCA) is proposed, and the influence of PCA on MCLP classification method is studied. One dataset, which comes from a bank in US, is used to test the performance of this approach, and the advantage of this classification method is shown by experiments