Machine Learning for Credit Card Fraud Detection
Yuxin Gao, Shuoming Zhang, Jiapeng Lu · 2021
With the development of E-bank, the use of credit cards gets an unprecedented improvement as well as the problem of credit card fraud. To overcome this problem, we need automatic systems to finish the fraud detection. The number of monitored account data is so large that our human resources are unable to detect the whole dataset. Also, since the number of fraudulent transactions is (fortunately) much smaller than the legitimate ones, the data distribution is unbalanced, skewed towards non-fraudulent observations. To solve the problem of unbalanced dataset, there exist many learning algorithms which are used to underperform this kind of problem; to improve the accuracy or predicting, there exist many methods like over or under sampling.