Restaurant Recommendation System using Machine Learning Algorithms

The necessity for recommendation systems spans across various domains, not least in the realm of dining and exploring new locales. Restaurants, in particular, benefit from such systems to both attract customers and offer them a chance to indulge in preferred and renowned dishes. Discovering these dishes, especially in unfamiliar areas, poses a significant challenge. This paper introduces a restaurant recommendation system that leverages food and service rating distributions to calculate matrix density. Additionally, we implement a popularity-based model to suggest restaurants to patrons. A ranking mechanism is utilized, based on scoring, to recommend top-rated restaurants and their most popular dishes. We use singular value decomposition and collaborative filtering to improve the model. The model’s performance is assessed using the RMSE metric. This study is conducted using data from Kaggle and includes the development of a web-based application created with Python’s Flask framework.