A Class Balancing Based Machine Learning Approach for Fraudulent Credit Card Transaction Detection
Gyana Ranjan Patra, Schaurya P. Iyer, Satyam Satyaprakash, Mihir Narayan Mohanty · 2023
The increased usage of credit cards for online and offline transactions in the past several years has also attracted criminals to commit fraudulent activities too. Since the volume and number of such transactions are increasing machine learning methods find a lot of applications in the detection of such anomalies. The data sets available for such cases are also highly imbalanced which add to the difficulty in the detection of such fraudulent activities. In this paper six machine learning algorithms namely Logistic Regression (LR), Multi-Layer Perceptron (MLP), Gradient Boosting Classifier(GBC), Random Forest Classifier(RFC), Decision Tree Classifier(DTC), Extreme Gradient Boosting Classifier(XGB) have been used along with a class balancing technique called Borderline SMOTE. The combination of class balancing with machine learning enhances the performance of all classifiers with the Extreme Gradient Boosting Classifier being the best.