Research And Application Of Trust-based Collaborative Filtering Recommendation Algorithm

With the rapid development of the Internet, there is a large amount of moviesresources in the network, which leads to the problem of information overload. It’s notefficient to find favorite movies using the traditional searching engine. If a userdoesn’t know anything about a movie, it’s difficult for him to find his favorite. Movierecommender systems, which discover users’ favorite movies by their preferences,can alleviate the problem of information overload greatly. Nowadays, there are twoweaknesses in the existing movie recommender systems. Firstly, traditionalcollaborative filtering recommendation technologies are widely used in therecommender systems. However, they suffer the inherent weaknesses, such as datasparseness problem, which results in that the recommender systems can’t make goodrecommendations when the data is extremely sparse. Secondly, movie recommendersystems haven’t incorporated the social sharing very well, which leads to a poorreal-time interaction.Based on these two problems, this thesis has finished some jobs. Firstly, I haveparticipated in the research of User Trust-based Collaborative FilteringRecommendation Algorithm (UTCF for short) as the third author, and this algorithmcan solve the data sparseness problem. Secondly, by utilizing the UTCF algorithm andincorporating the social sharing, this thesis has designed and developed a system toshare and recommend movies, which can explore favorite ones according to thepreferences of the user.This system has some functions. Firstly, it can recommend movies which users are interested in efficiently in a mass of information, and provide valuable suggestionsfor users in choosing them. What’s more, it can recommend friends with whom usershave similar taste. Users can share the comments with friends and find their favoritemovies.According to the movie recommender systems and social sharing, this thesis hasanalyzed the requirement of the system. Then, based on the feature of the UTCFalgorithm, the system has been designed in detail. At last, it has been developed withthe technologies of Struts, Spring and Hibernate. The feature of this system is that theelement of social sharing is integrated into the movie recommender system smoothly.It utilizes the recommender system to recommend movies and friends, and shows thestrong interactivity character of social sharing. Not only can it help users find favoritemovies efficiently, but also it achieves the economic value of movie promotion.