Back to all selected work
Applied Machine Learning · Research·2025·research·Best Presentation Award — IEEE ICCPCT 2025
Credit Card Fraud Detection ML
Anomaly detection research published in IEEE Xplore for ICCPCT 2025.
PythonScikit-LearnXGBoostImbalanced-LearnIEEE Xplore
Overview & Context
Research paper and benchmark models addressing severe class imbalance in real-time transaction fraud scoring.
The Problem
Fraudulent transactions typically account for less than 0.2% of stream volume, causing standard classifiers to suffer high false-negative rates on critical breaches.
The Architectural Approach
Evaluated hybrid sampling techniques (SMOTE, NearMiss, Random Under-Sampling) combined with tuned XGBoost and Random Forest ensembles. Published in IEEE Xplore for ICCPCT 2025 and awarded Best Presentation.
Key Technical Highlights
- Published in IEEE Xplore for ICCPCT 2025
- Awarded Best Presentation Award at IEEE ICCPCT
- Comprehensive benchmark of cost-sensitive sampling and gradient boosted trees
Next Case Study
View projectFRIP — Real-Time Fraud Scoring
Streaming transaction fraud scoring on Kafka with ML anomaly detection.