A recommendation system for online social semantic network using knowledge based, content based and collaborative filtering

Abstract

Recommendation systems are a very popular service whose accuracy and sophistication keeps increasing every day. Yet current systems pose a limitation on personalized user recommendation, which we wish to improve. We are developing Content-Based, Collaborative Filtering and Knowledge-Based models and we wish to find the most appropriate approach to build restaurant recommendation systems. We followed steps that involved a pipeline to process reviews of restaurants obtained from a widely used online network of zomato users (India’s largest restaurant service) and calculate ratings of restaurants from reviews. Using a machine learning technique, it continuously analyses user restaurant visit patterns.