Optimizing Credit Card Fraud Detection Using a Hybrid Sampling Approach with SVM

Credit card fraud detection remains a significant challenge due to the highly imbalanced nature of transactional datasets, where fraudulent transactions represent only a tiny fraction. This paper presents a hybrid sampling approach that integrates Adaptive Synthetic Sampling (ADASYN), Synthetic Minority Oversampling Technique (SMOTE), and Random Undersampling to address this imbalance, using a Support Vector Machine (SVM) classifier. The proposed method was evaluated using a real-world credit card dataset and compared to traditional resampling techniques. Results demonstrate that the hybrid approach significantly improves the detection of fraudulent transactions. The SVM model trained with SMOTE achieved an Fl-score of 0.9290 and an AUC-ROC score of 0.9306, slightly outperforming the ADASYN-based model. This study highlights the effectiveness of combining different resampling techniques to enhance fraud detection in imbalanced datasets.