Credit Card Fraud Detection using Machine Learning Algorithms

Abstract

Due of digitalization and expansion of the Internet business, there has been an increment in online exchanges and the credit card transactions sum for an enormous piece of these exchanges. As a result of this, the credit card frauds being committed has increased immensely. Credit card frauds cause huge losses for customers as well as the banks and financial institutions and hence have to be dealt with carefully and quickly. It is important that the credit card frauds are detected with an accuracy of 100% so that the customers are not charged for items that they didn’t purchase. Therefore, credit card fraud detection applications have been given much greater importance these days. This problem can be solved using Data Science along with Machine Learning. In this paper, we have used Machine Learning algorithms like Local Outlier Factor, Isolation Factor, Random Forest and K-Nearest Neighbor on the credit card transaction dataset to find fraudulent transactions. The efficiency of these algorithms is assessed based on the performance metrics like accuracy, precision, recall, f1-score and support and the most efficient algorithm is identified.