Selection of optimal credit card fraud detection models using a coefficient sum approach

Abstract

During the past years various fraud detection techniques were proposed and used to reduce fraudulent activities. Selection an optimal fraud detection model becomes keen area of interest for researchers in the field of anomaly detection. Methods and tools for fraud detection model selection which were used in the literature used a limited no of model selection criteria for finding the detection of fraud in various areas. So for the first time we propose a approach based on a Coefficient sum method. This approach is used for ranking of Fraud detection models for finding the optimal model from various fraud detection models. For ranking the FDMs, various model selection criteria, with a set of FDMs are require Real data sets are used for illustration of the Coefficient sum method. The result of this paper gives a ranking method based on the Sum value of each criteria of FDM.