Credit Card Fraud Detection using Machine Learning
Anjali Singh Rathore, Ankit Kumar, Depanshi Tomar, Vasudha Goyal, Kaamya Sarda, Dinesh Vij
Abstract
Credit card fraud is a common and fast-growing issue. Such problems may be tackled with Data Science, which, together with Machine Learning, should not be overlooked. This paper compares the performance of Decision Tree, Random Forest, K-nearest neighbors, and Logistic regression on highly imbalanced data. This is done by merging all important features of cardholder transactions, such as the date, user zone, product category, amount, supplier, and client’s behavioral habits, among others. The information is then put into different models that look for patterns and rules to evaluate whether or not a transaction is fraudulent based on accuracy and sensitivity.