SECURITY PROVISION FOR UPI TRANSCATIONS USING MACHINE LEARNING TO DETECT FRAUD TRANSCATIONS

ABSTRACT: This project presents a comprehensive fraud detection system using machine learning techniques to identify fraudulent transactions from financial data. The dataset undergoes extensive preprocessing, including column renaming, data type conversion, and feature extraction. Exploratory Data Analysis (EDA) is performed using visualizations to uncover trends and anomalies in the data. Various models, including Decision Trees and XGBoost, are applied to classify transactions as fraudulent or legitimate. Additionally, Synthetic Minority Over- sampling Technique (SMOTE) is used to handle class imbalance, and Principal Component Analysis (PCA) is applied to reduce dimensionality. Performance evaluation using accuracy, confusion matrix, and classification reports demonstrates the effectiveness of the proposed fraud detection system. The model is further saved for deployment using pickle for practical use. KEYWORDS: Fraud Detection, Machine Learning, XGBoost, Decision Tree, SMOTE, PCA, Financial Data Analysis, Data Preprocessing, Classification, Python.