Optimising credit card fraud detection through machine learning and deep learning with spatial-temporal imbalance handling
- Publication Type:
- Thesis
- Issue Date:
- 2024
Open Access
Copyright Clearance Process
- Recently Added
- In Progress
- Open Access
This item is open access.
The rapid increase in financial transactions conducted online has intensified dependence on digital payment systems and raised financial fraud, highlighting the need for effective fraud detection systems. This research addresses the "class imbalance challenge" in credit card fraud detection by integrating geolocation and temporal analysis to improve trend identification and anomaly detection. We devised an innovative methodology using sophisticated machine learning (ML) and deep learning (DL) approaches in conjunction with data balancing techniques, including random over sampling (ROS), synthetic minority over-sampling technique (SMOTE), adaptive synthetic sampling (ADASYN), and random under sampling. We assessed eight machine learning algorithms—Bagging Classifier, Random Forest, CatBoost, Logistic Regression (LR), Extreme Gradient Boosting (XGBoost), AdaBoost, Gaussian Naive Bayes, and Extra Trees Classifier—and two deep learning models, Gated Recurrent Unit (GRU) and Neural Network (NN). Performance was evaluated using Recall, Precision, F1 Score, ROC-AUC Score, and Accuracy. The Bagging Classifier and Random Forest Classifier models exhibit exceptional performance, achieving impressive results across all metrics. This demonstrates their capacity to effectively identify fraudulent transactions while keeping the rate of false positives to a minimum. This comprehensive, multi-faceted strategy effectively responds to the complexities of digital financial fraud.
Please use this identifier to cite or link to this item: