Related Experiment Videos
Applying supervised machine learning algorithms and ensemble models to enhance credit card fraud detection
Abrar Al-Bulushi1, Abdul K Shaikh2, Naresh Adhikari3
1Sohar International Bank, Muscat, Oman.
Frontiers in Artificial Intelligence
|June 10, 2026
Summary
This study enhances credit card fraud detection using machine learning and ensemble methods. The best approach combined bagging ensemble techniques with resampling for improved accuracy and robustness in identifying fraudulent transactions.
Area of Science:
- Computer Science
- Data Science
- Financial Technology
Background:
- Credit card fraud poses a significant economic threat globally.
- Sophisticated fraudulent activities necessitate advanced detection systems.
- Existing methods struggle with highly imbalanced financial datasets.
Purpose of the Study:
- To improve credit card fraud detection accuracy and model robustness.
- To explore the efficacy of machine learning and ensemble learning techniques.
- To develop a scalable framework for real-world imbalanced financial data.
Main Methods:
- Implemented supervised learning algorithms: Decision Tree, Logistic Regression, Naïve Bayes, Random Forest, Artificial Neural Network, and XGBoost.
- Applied resampling techniques (RUS, ROS, SMOTE) to address data imbalance.
- Utilized ensemble methods (bagging, boosting, stacking) and feature selection.
Main Results:
- Ensemble learning models significantly outperformed standalone algorithms.
- The bagging ensemble technique achieved 0.99 accuracy, 0.90 recall, and 0.77 precision.
- The best model combined Decision Tree, Random Forest, and Artificial Neural Network with resampling, showing strong generalization.
Conclusions:
- Integrating ensemble learning, resampling, and diverse machine learning models substantially enhances fraud detection.
- The proposed framework offers a robust and scalable solution for detecting credit card fraud.
- The approach maintains high predictive performance on imbalanced financial datasets.