Movie Recommendation Systems Based on Collaborative Filtering: A Case Study on Netflix

Abstract

User ratings on items like movies, songs, and shopping products are used_x000D_ by Recommendation Systems (RS) to predict user preferences for items that have_x000D_ not been rated. RS has been utilized to give suggestions to users in various domains_x000D_ and one of the applications of RS is movie recommendation. In this domain, three_x000D_ general algorithms are applied; Collaborative Filtering that provides prediction_x000D_ based on similarities among users, Content-Based Filtering that is fed from the_x000D_ relation between item-user pairs and Hybrid Filtering one which combines these_x000D_ two algorithms. In this paper, we discuss which methods are more efficient in movie_x000D_ recommendation in the framework of Collaborative Filtering. In our analysis, we use_x000D_ Netflix Prize dataset and compare well-known Collaborative Filtering methods_x000D_ which are Singular Value Decomposition, Singular Value Decomposition++, KNearest Neighbour and Co-Clustering. The error of each method is calculated by_x000D_ using Root Mean Square Error (RMSE). Finally, we conclude that K-Nearest_x000D_ Neighbour method is more successful in our dataset.