Credit card Fraud Classification using an Optimized Ensemble Learning Technique

Recent advancements in e-payment and e-commerce methods have resulted in rising the quantity of credit card transactions that are fraudulent, which cause significant massive financial losses and become a potential security issue. There is an urgent need for efficient methods for identifying fraudulent credit card transactions. This paper presents an effective ensemble learning technique that utilizes the grid search optimization approach for identifying credit card fraud. The suggested approach consists of two phases. First, base learners consist of multiple machine learning classifiers, including Decision Tree (DT), K-nearest neighbor (KNN), AdaBoost (ADA), Gradient Boosting (GB) and Logistic Regression (LR), are utilized to find the fraudulent transactions probabilities. Second, a meta learner that integrates the Random Forest with the Grid Search (RF-GS) is applied to categorize the probabilities of predictions produced by the base learners. RF-GS uses the Grid Search (GS) optimization technique to tune the parameters of Random Forest (RF) method, to get the maximum credit card fraud detection accuracy. A real-world dataset was utilized to evaluate the effectiveness of the suggested approach. The findings of the experiment show the effectiveness of the suggested optimized ensemble-learning strategy for identifying the fraudulent credit card transactions, which performed better than the other approaches and obtained superior accuracy of 99.01%.