Collaborative Filtering Recommendation System Based Upon User Reviews

摘要

Nowadays, there are many recommendation systems, accessible via internet, which attempt to recommend to users several products such as music, movies, books, etc. Aiming at long response time and solving cold start problems that are faced by present recommendation algorithm. This paper, proposes a collaborative filtering approach based on user's credibility taking Movies as an example. This approach will find out the cluster that target user belongs to and further provide recommendation. Collaborative model will improve the response time, increased the performance and find out the Mean Absolute Error. Section first describes Introduction about Collaborative recommendation system, its work flow and why it is used. In second section related work about Collaborative Filtering. Section third describes how to find out the Mean Absolute Error and how to reduce it by using the Euclidean distance and Pearson correlation. And at last in forth section its experimental evaluations to predict the MAE and time to build the recommendation for each user.