TourOptiGuide: A Hybrid and Personalized Tourism Recommendation System

Abstract

When visitors explore a city briefly, they must prioritize the key attractions that align with their interests. These significant points of interest (POIs) can be chosen based on specific criteria tailored to their needs. Additionally, travellers venturing into unfamiliar regions often seek help to plan their itinerary. To address this issue, we developed and presented a novel hybrid and personalized recommendation system aimed at helping tourists choose their next POI. The system tailors its suggestions based on four key factors: the tourist's current location, single preferences, age, and historical experiences. Deep learning models play a crucial role in identifying the tourist's current location from images and predicting age from selfies. In addition, our system leverages a trajectory data warehouse containing extensive historical data of past tourist’s experiences to provide suggestions. The core of our recommendation strategy is a fuzzy logic decision support system. This system effectively synthesizes diverse inputs to produce the top next POI to visit. By integrating various recommendation methods, our hybrid system significantly improves the precision and pertinence of its recommendations, offering a more customized and effective travel experience. Preliminary results demonstrate significant improvements in tourist satisfaction and in the efficiency of itinerary planning.