DESIGN AND IMPLEMENTATION OF CREDIT CARD FRAUD DETECTION ON WEB APP USING MACHINE LEARNING (LOGISTIC REGRESSION)

ABSTRACT

This project is aimed at developing a reliable user-friendly web-based credit card fraud detection system using machine learning (logistic regression). The problem involved inadequacy of existing fraud detection systems, leading to significant financial losses and decreased trust in banking systems will be addressed in carrying out by the researched work. The methodology involved data collection, preprocessing, model training, and evaluation, as well as web application development with a focus on user interface and security. The result demonstrates the development of a user-friendly online interface for a credit card fraud detection system, emphasizing simplicity and ease of use. Key features include a straightforward method for uploading transaction data and clear user guidance. The system effectively differentiates between normal and fraudulent transactions, validated by its accurate performance. The dataset used for training machine learning models included over 284,807 transactions, successfully distinguishing 492 fraudulent transactions from 284,315 legitimate ones. The conclusion highlights the system's effectiveness in accurately detecting fraudulent transactions while minimizing false positives and false negatives, with recommendations for ongoing updates, security enhancements, and user education to address evolving fraud patterns and cyber threats.