Restaurant Recommendation System Based on ML Algorithms and Real-Time Web Scraping

As the restaurant industry expands, the sheer number of dining options makes it increasingly difficult for individuals to find establishments that match their preferences. To address this challenge, we present a personalized restaurant recommendation system that leverages both collaborative filtering and content-based techniques, augmented by real-time web scraping. Users provide their past restaurant ratings and desired location, which is used to recommend dining options based on two algorithms: user similarity and Google reviews. If a location is absent in the database, the system dynamically scrapes Google reviews and applies natural language processing (NLP) to extract feature-opinion pairs, identify sentiment, and compute a reliability score. Collaborative filtering identifies user similarities based on historical ratings, while the content-based model combines sentiment analysis, average ratings, review count, and reviewer reliability to generate a comprehensive restaurant score. Recommendations are ranked and accompanied by a detailed breakdown of the contributing factors, offering users transparency in how the suggestions are tailored to their preferences. This approach not only enhances the accuracy of recommendations but also provides a more informed and personalized dining experience.