Enhancing Restaurant Recommendations through User-Based Collaborative Filtering

In this study, a restaurant recommendation system based on user-based collaborative filtering was rigorously evaluated using the Yelp dataset. We developed a comprehensive methodology that involved data loading, matrix transformation, the normalization of user ratings and finding similarity. Overall, this study presents a disciplined approach to constructing a recommendation system that enhances user experiences by aligning suggestions with preferences, promising further advancements in restaurant recommendation systems and contributing to the culinary industry’s growth and enrichment. Our evaluation, measured by Root Mean Square Error (RMSE), revealed that considering the preferences of the top 5 nearest neighbors yielded the most accurate recommendations, highlighting the system’s robustness. The RMSE value of top 5 is 0.134096542. Interestingly, the system’s accuracy remained consistent across various neighbor-counting scenarios, ensuring users consistently received reliable restaurant suggestions.