Font Size: a A A

Research And Implementation Of Real-time Movie Recommendation System Based On Spark Streaming Computing

Posted on:2019-09-25Degree:MasterType:Thesis
Country:ChinaCandidate:L YanFull Text:PDF
GTID:2428330572463622Subject:Computer technology
Abstract/Summary:PDF Full Text Request
Nowadays,with the rapid development of science and technology,the amount of information on film is increasing day by day.The information overload related to film has become a problem that can not be ignored.When we are ready to watch a movie to relax,it may be difficult to find a satisfying film to relax because there is too much information about the film.In addition,a large number of users' implicit feedback has not been used,so how to use implicit feedback from the vast amount of movie information to find their own want to see the movie has become a research hotspot.In addition,there is no recommendation for new users and new movies because of lack of relevant data.In view of the above problems,this paper designs and implements a real-time movie recommendation system based on Spark streaming computing.In order to combine the implicit feedback of users with the display feedback of users,the real-time movie recommendation for users is carried out.It also solves the problem that new users and new movies can not be recommended in time.The main research work in this paper is as follows:1.build off line computing system based on model based collaborative filtering algorithm on Spark platform.A cold start method is designed to solve the matrix decomposition collaborative filtering algorithm.Compared with the original collaborative filtering algorithm,the recommendation accuracy is improved.2.build an online computing system using Spark Steraming on the Spark platform.An analog log program is designed,which can generate user log files in real time;a Spark Streaming real-time receiving module is designed to receive and process user log files in real time;and how to quantify user implicit feedback to compute attention is proposed,and an online recommendation list is obtained by combining the attention with the off-line algorithm model.3.use java to design the WEB side of the movie recommendation system.User operation browse module and administrator operation management module aredesigned.User operation browsing module can provide users with registration,real-time recommendation and other functions;Administrator operation management module provides film addition and Off-Shelf functions.
Keywords/Search Tags:Recommendation System, The Movie, Collaborative Filtering, log file, Spark Streaming
PDF Full Text Request
Related items