FINANCIAL FRAUD DETECTION

Abstract: This project presents a machine learning-based financial fraud detection system designed to enhance the security and reliability of digital financial transactions. The model analyzes key transactional and behavioral features such as transaction amount, time, location, customer spending patterns, and account history to accurately detect fraudulent activities. Multiple classification algorithms were evaluated, including Logistic Regression, Support Vector Classification (SVC), Decision Tree, Random Forest, and Multilayer Perceptron (MLP). Among these, the Random Forest algorithm achieved the highest accuracy of 98.9%, demonstrating superior capability in handling imbalanced and complex financial datasets. The system was deployed as an interactive web application using Streamlit, enabling real- time fraud prediction and alert generation. This work highlights the potential of ensemble and deep learning approaches for secure, data-driven financial systems, offering an efficient and scalable solution to mitigate fraud risks and enhance transaction safety.

Keywords: machine learning, financial fraud detection, random forest, anomaly detection, decision tree, SVM, logistic regression, multilayer perceptron.