A Credit Card Fraud Model Prediction Method Based on Penalty Factor Optimization AWTadaboost  

作　　者：Wang Ning Siliang Chen Fu Qiang Haitao Tang Shen Jie 

机构地区：[1]College of Computer and Communication,Hunan Institute of Engineering,Xiangtan,411104,China [2]College of Computational Science and Electronics,Hunan Institute of Engineering,Xiangtan,411104,China

出　　处：《Computers, Materials & Continua》2023年第3期5951-5965,共15页计算机、材料和连续体（英文）

基　　金：This research was funded by Innovation and Entrepreneurship Training Program for College Students in Hunan Province in 2022(3915).

摘　　要：With the popularity of online payment, how to perform creditcard fraud detection more accurately has also become a hot issue. And withthe emergence of the adaptive boosting algorithm (Adaboost), credit cardfraud detection has started to use this method in large numbers, but thetraditional Adaboost is prone to overfitting in the presence of noisy samples.Therefore, in order to alleviate this phenomenon, this paper proposes a newidea: using the number of consecutive sample misclassifications to determinethe noisy samples, while constructing a penalty factor to reconstruct thesample weight assignment. Firstly, the theoretical analysis shows that thetraditional Adaboost method is overfitting in a noisy training set, which leadsto the degradation of classification accuracy. To this end, the penalty factorconstructed by the number of consecutive misclassifications of samples isused to reconstruct the sample weight assignment to prevent the classifierfrom over-focusing on noisy samples, and its reasonableness is demonstrated.Then, by comparing the penalty strength of the three different penalty factorsproposed in this paper, a more reasonable penalty factor is selected.Meanwhile, in order to make the constructed model more in line with theactual requirements on training time consumption, the Adaboost algorithmwith adaptive weight trimming (AWTAdaboost) is used in this paper, so thepenalty factor-based AWTAdaboost (PF_AWTAdaboost) is finally obtained.Finally, PF_AWTAdaboost is experimentally validated against other traditionalmachine learning algorithms on credit card fraud datasets and otherdatasets. The results show that the PF_AWTAdaboost method has betterperformance, including detection accuracy, model recall and robustness, thanother methods on the credit card fraud dataset. And the PF_AWTAdaboostmethod also shows excellent generalization performance on other datasets.From the experimental results, it is shown that the PF_AWTAdaboost algorithmhas better classification performance.

关 键 词：Credit card fraud noisy samples penalty factors AWTadaboost algorithm 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]