Hybrid Recommender System: Recommending Restaurants to Users

Abstract

The objective of the project is to recommend restaurants to users. We built different collaborative and content-based recommendation models and evaluated their performances using Yelp Dataset. For content-based approaches, we created business and user profiles and computed similarity between them to make personalized recommendations. We evaluated and compared the performance of the algorithms using RMSE (Root Mean Squared Error) metric. We developed a hybrid model by combining all the individual models which improved the overall RMSE.