A HYBRID BLOCKCHAIN AI FRAMEWORK FOR SECURE AND FRAUD-RESISTANT TRAVEL AND HOSPITALITY BOOKING TRANSACTIONS

Abstract

The study investigates the development and testing of a hybrid blockchain-artificial intelligence system that secures travel and hospitality booking platforms from fraudulent transactions. The research combines three technological aspects: an immutable Hyperledger blockchain ledger to ensure tamper-proof transactions are recorded, an ensemble AI fraud detection model that consists of XG Boost, Random Forest, and Isolation Forest classifiers, and decentralized self-sovereign identity verification using Hyperledger indy with zero knowledge proofs.  The framework was validated using the AI Hotel Marketplace Fraud Detection dataset consisting of 791 OYO Rooms bookings from major Indian cities.  The ensemble AI model achieved better results through 96.6% accuracy together with 0.806 F1-score because it outperformed single classifiers by 9.4%. Price-related variables showed the highest predictive power while geographic analysis found fraud concentrated in particular areas which had fraud rates that neared 18%. The blockchain layer-maintained transaction unchangeability through its ability to verify hashes within sub-second intervals while decentralized identity systems protected user data by 67% through controlled data sharing which required 1.8-second verification times. The research introduces a novel integrated blockchain-AI-identity framework which enables secure travel booking systems by tackling cross-jurisdictional transaction security issues. The framework meets U.S. national cybersecurity standards through its implementation of zero-trust architecture and National Artificial Intelligence Initiative Act requirements for reliable AI systems.