Credit Card Fraud Detection Using Machine Learning

This study investigates credit card fraud detection using various machine learning techniques. It emphasizes handling the imbalanced nature of fraud data through a combination of oversampling (10%) and undersampling (50%) methods. The most effective model, Random Forest, was identified using a comparison of different sampling techniques. The findings demonstrate the efficacy of integrating sampling techniques for improved model performance in fraud detection, offering insights into optimal strategies for addressing class imbalance in financial datasets.