Hybrid Approach Using Machine Learning Techniques in Credit Card Fraud Detection

Abstract

Crime financial services have a serious problem by credit card fraud. Every year billions of dollars are lost by credit card fraud. Lack of research and studies on analyzing real-world credit card data is there. In this paper, we apply machine learning algorithms to detect the credit card fraud using credit card datasets and compare their performance. We use hybrid algorithm model involving voting classifier to detect the fraud. The result positively indicates that the voting classifier method gives good accuracy, precision, recall, and F1 score in detecting fraud in credit card transactions. More than that, the work identifies the important variables in the dataset that may lead to higher accuracy in credit card fraud detection, and on the whole, we discuss about various types of classifiers and its performance in detecting fraudulent transactions.