Comparative Study of Machine Learning Based Classification Techniques for Credit Card Fraud Detection

Abstract

We live in the digital era where paying by credit card has become an essential part of our life. With the increasing demand for credit cards, it has also become a challenge to maintain the integrity of users. Credit card fraud is a malignant activity that causes financial loss that costs consumers as well as banks. Nowadays, fraudsters use distinct methods to commit fraud. To identify credit card fraud, it is important to understand the type of fraud and past transaction history. To solve credit card-related issues there should be a system for fraud detection that can detect fraud activities before they occur and also in an accurate way. In this paper, we implement six widely used machine learning techniques for credit card fraud detection. For each machine learning technique, a confusion matrix is prepared for performance analysis of the algorithm. Their efficacy is analyzed based on the parameters such as accuracy, precision, recall, specificity, misclassification, and F1 score. Results show that machine learning techniques are helpful for credit card fraud detection. Although results show a high level of precision and recall for each algorithm, we strongly recommend using multiple machine learning techniques for fraud detection before they occur or in the process of occurrence.