A comprehensive and relative study of detecting deformed identity crime with different classifier algorithms and multilayer mining algorithm

Abstract: No w a day’s identity crime is well known, prevalent and prominent in our society. There are some algorith ms wh ich have been implemented to detect or resolve resilience identity. The existing data mining and non data mining algorith m and known fraud matching have some limitation. To achieve a co mp lete and transparent view of different resilience identity crime detection system a comparative study and approach is a prime focus of this paper. Here different classifiers algorith m are co mpared with data min ing based algorithm with mult ilayer min ing stage of defense (territory detection and suspicion score detection algorithm) to probe the synthetic identity crime. Through this comprehensive approach we can find out the relative measure ment of different classifier algorith m using proposed algorith m territory detection and suspicion score detection algorithm as baseline. A lthough multilayer min ing algorith m is specific to cred it application fraud detection, but the concept of resilience o r deformat ion with this comparat ive study discussed in this paper are general to design, imp lement and evaluate of all detection system.

Keywords: Data mining based fraud detection, security, anomaly detection, data stream min ing.