A Data Mining Based System For Transaction Fraud Detection
Deng Wenkai, Ziming Huang, Jiachen Zhang, Junyan Xu
Abstract
With the deepening of world trade informationization degree, transaction fraud has been endangering the security of world finance and commerce. The frequency and scale of transaction fraud are expanding day by day, which makes the vast number of users and financial practitioners suffer huge economic losses. With the increasing maturity of data mining and machine learning in the field of computer science, the detection of transaction fraud gradually finds a practical solution. This paper adopts a transaction fraud detection system based on random forest and manual detection. The experimental results of IEEE CIS fraud dataset show that the method of this model is better than the benchmark model, such as logistic regression, support vector machine. Finally, the accuracy of our model reached 96.8%, and the AUC ROC score was 92.5%.