Research On The Personalized Movie Recommendation System Based On Hybrid Collaborative Filtering

Along with the rapid development of Internet,the information on the Internet are growing exponentially,the problem of “information overload” is becoming more and more serious.The recommendation system provide services by constructing the user interest model,and has effectively alleviated the information overloading problem.As one of the main forms of information acquisition and cultural consumption,movies have become one of the channels for people to pursue spiritual culture.Usually it is difficult for people to accurately express their needs.Secondly,the movies on the Internet are increasing constantly.This makes the contradiction between the massive supply of movies and the diversified demands of users.At this time,the organic combination of personalized recommendation technology with the movie system can address the imbalance between movie supply and demand.A personalized movie recommendation system can not only help users find their favorite movies quickly,but also enhance the loyalty to the movie recommendation system.This can help the operators to increase their economic returns with VIP services,advertising,and other means.This paper takes the personalized movie recommendation system as the research object,studies the algorithm and application of the personalized recommendation technology.This paper proposes a personalized movie recommendation algorithm based on hybrid collaborative filtering,and validates its effectiveness on the Movielens data set.Then designes and implements a personalized movie recommendation system based on hybrid collaborative filtering.With this aim,the main contents are as follows:(1)Summarizing relevant literature from the aspects of collaborative filtering algorithm and personalized movie recommendation,then pointing out the deficiencies of the current research.(2)A common scoring weight function is introduced to improve the overestimation of similarity due to data sparsity problem.A time weight function which gives a time weight to score at different times is introduced to improve the calculation method of score prediction,solving the timeliness of movie recommendation effectively.The experimental results shows that the improved collaborative filtering algorithm has better recommendation accuracy.(3)Based on the improved collaborative filtering algorithm,a hybrid collaborative filtering recommendation algorithm based on usermovie is proposed,and the main idea is to mix the predicted scores from user-based and movie-based collaborative filtering by a linear weighted method.In terms of the similarity weights of the user's and movie's neighbor sets,the paper decomposes the weight factors.The experimental results shows that the hybrid collaborative filtering algorithm is better on recommendation accuracy.(4)From practical aspect,implementing a recommendation engine based on a hybrid collaborative filtering algorithm and constructing a personalized movie recommendation system based on B/S architecture,which is to achieve functions such as movie searching,movie scoring,movie recommendation,individual center etc.