AI-Powered Fraud Detection in Financial Transactions

Abstract

Fraud detection in financial transactions is a critical task that this study addresses using machine learning models like MLP, XGBoost, and LSTM. By applying data preprocessing, feature engineering, and model training on historical transaction data, the approach analyzes key features such as transaction amount, frequency, location, and user behavior to distinguish between legitimate and fraudulent activities. MLP captures complex patterns, XGBoost boosts accuracy through ensemble learning, and LSTM detects sequential fraud effectively. Evaluation using accuracy metrics shows that this AI-powered method enhances real-time fraud prevention, reduces false positives, and strengthens overall financial security.