A Random Forest and Logistic Regression Based Tool for Credit Card Fraud Detection

Abstract

Detecting credit card fraud is an important issue in financial transactions since it could result in large financial losses and data breaches. However, traditional rule-based systems frequently struggle with imbalanced datasets and are unable to adapt to changing fraud strategies. This project suggests a machine learning-based fraud detection tool that identifies transactions as either valid or fraudulent by using Random Forest and Logistic Regression. The Synthetic Minority Oversampling Technique (SMOTE) is employed to handle data imbalance and enhance model accuracy. Implemented as a web-based platform with Flask, the tool allows fraud analysts in banking and financial institutions to upload transaction files, analyse results, and visualize fraud patterns. The tool is expected to address the limitations of traditional methods and offer an approach for securing digital financial systems by decreasing false positives and improve fraud detection accuracy.