A Secure and Scalable Online Auction System Leveraging Hybrid Machine Learning Models and Rule-Based Systems for Real-Time Fraud Detection and Transparent Bidding

Abstract

The Web-Based Online Auction Platform is an innovative e-commerce solution designed to facilitate secure and dynamic transactions between buyers and sellers. Supporting multiple auction formats, including Silent Auction, Buy-It-Now, and Standard Auction, the platform ensures flexibility and engagement for users. Real-time bidding updates are seamlessly integrated using Firebase, enhancing interaction and responsiveness. To strengthen security and trust, the system incorporates advanced Machine Learning (ML) and Deep Learning (DL) models, such as Random Forest, Isolation Forest, and Autoencoders, for fraud detection. These models analyze bidding behaviors, identify anomalies, and mitigate risks associated with fraudulent activities. For transaction security, the platform leverages Razor pay-powered escrow payments, ensuring that funds are securely held until both parties confirm the transaction, thereby minimizing disputes and enhancing user confidence. The platform is designed with a modular architecture, ensuring scalability and efficient handling of large user bases. This research highlights the system’s capability to offer a secure, interactive, and user- centric online auction experience with robust fraud detection mechanisms and transaction security features, making it an ideal solution for organizations and individuals seeking a reliable digital auction environment.