Improvement in credit card fraud detection using ensemble classification technique and user data

Abstract

Financial fraud a serious problem in banking system. Credit card fraud growing with increasing Internet usage, as it becomes very simple to collect user data and do fraud transaction. Fortunately, all records including fraud and legit transactions are present in the financial record. Improved data mining techniques are now capable to find solutions for such outlier detections. Financial data freely available in many sources, but this data has some challenges like,1) the profile of legit and fraudulent behavior changes constantly, 2) there a class imbalance problem in dataset, because less than 3% transaction are fraud, 3) transaction verification latency also one more problem. All this data issues are handled using pre-processing techniques like cleaning and reduction. Main aim of this research to find out, output attribute is Fraud, with better time complexity. To this end, K-means, Random Forest and J48 algorithm used, and its accuracy rates are compared to find best fit pre-processing and machine learning algorithm. It observed that accuracy rate of Random Forest 93.8% when both global and local dataset used.