Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This paper introduces a machine learning framework for fraud detection in banking transactions using the PaySim dataset. It addresses class imbalance with SMOTE and optimizes model performance via GridSearchCV-based hyperparameter tuning of Logistic Regression, Random Forest, XGBoost, and Decision Tree classifiers. Results demonstrate a robust and scalable solution for enhancing fraud prevention in FinTech.
Hyperparameter-optimized XGBoost on SMOTE-balanced data significantly boosts fraud detection, offering a practical defense against increasingly sophisticated financial attacks.
The expansion of digital payment systems has heightened both the scale and intricacy of online financial transactions, thereby increasing vulnerability to fraudulent activities. Detecting fraud effectively is complicated by the changing nature of attack strategies and the significant disparity between genuine and fraudulent transactions. This research introduces a machine learning-based fraud detection framework utilizing the PaySim synthetic financial transaction dataset. Following the CRISP-DM methodology, the study includes hypothesis-driven exploratory analysis, feature refinement, and a comparative assessment of baseline models such as Logistic Regression and tree-based classifiers like Random Forest, XGBoost, and Decision Tree. To tackle class imbalance, SMOTE is employed, and model performance is enhanced through hyperparameter tuning with GridSearchCV. The proposed framework provides a robust and scalable solution to enhance fraud prevention capabilities in FinTech transaction systems. Keywords: fraud detection, imbalanced data, HPO, SMOTE