Credit Card Fraud Detection Using Minority Oversampling and Random Forest Technique

In an era of digitalization, the usage of credit card is rapidly increasing and as a result, the cases associated with credit card fraud are on raise. So, fraud detection has become an important instrument and, in many cases, the best strategy to prevent fraud. Among the existing techniques, Machine Learning (ML) is crucial in detecting fraud. In this paper, ML algorithms like Decision Tree (DT), k Nearest Neighbors (KNN), Logistic Regression (LR) and Random Forest (RF) were analyzed and it is inferred that the RF algorithm works best on performance measures like precision, accuracy, recall, Matthews Correlation Coefficient, F-1 score. This paper also proposes to detect 100% fraud transactions while reducing false negatives by using SMOTE (oversampling technique) in combination with Random Forest algorithm.