AI-ENHENCED ROOM RENTAL SEARCH AND RECOMMENDATION SYSTEM

Abstract

Properties play an essential part in modern living. Our everyday interaction with rental properties is not just about viewing or renting them; behind each listing lies a story encapsulated within a complex digital ecosystem. With the widespread use of smartphones, people increasingly photograph and explore properties online. However, images alone do not convey the full rental process, nor do they effectively connect tenants and landlords.Therefore, in this thesis, we introduce an intelligent rental platform that facilitates property discovery, matching, and communication through an AI-assisted system. Our platform predicts property–tenant compatibility using a novel architecture capable of capturing inter-feature relationships without requiring a predefined structure. It then generates personalized property recommendations by jointly analyzing both property data and user preferences.Leveraging a curated dataset of over 15,000 rental listings containing images, amenities, and behavioral interaction data, our AI recommendation engine achieves an accuracy of 85% in predicting suitable property matches. For real-time interaction, the Socket.IO-based messaging module maintains a response latency below 200ms, demonstrating its efficiency in supporting seamless communication between tenants and landlords.This study, therefore, contributes meaningfully to the field of artificial intelligence and web technologies, particularly in advancing methods for property recommendation, AI-powered search, and real-time communication within rental platforms.