Design And Implementation Of Movie Recommendation System Based On Hadoop

With the development of society and the rapid innovation of technology,people gradually stepped into the period of information surplus from the age of inadequate consultation,in such an era,whether it is information producers or the users of information are facing unprecedented challenges: first,for the former,it is not an easy process to make the information they give different and get a lot of attention at the same time;second,For the latter,it is also not an easy process to find a way to benefit their needs from vast amounts of information,so the number of movies that consumers can watch on the Internet today is increasing in geometric multiples.In order to make the user experience better,personalized recommendation technology is an effective solution.In recent year the number of users and movies has picking up and the intranet information has rapidly expanding.The traditional recommendation algorithm is given priority to with the stand-alone and more complicated,The traditional recommendation algorithm can't meet the huge amount of date to recommend calculation.In order to make the recommendation algorithm more extensible and effective,The movie recommendation system in the paper choose the HDFS which has highly fault-tolerant distributed file system as the underlying file system,at the same time,we choose MapReduce as the tool of dealing with the abundant date.The paper identifies three priority recommendation algorithm,including Recommendation algorithm based on content,Recommendation algorithm based on association rules and recommendation algorithm based on collaborative filtering.Based on calculation and analysis,we found that the collaborative filtering algorithm based on project can better reflect the needs of users and improve efficiency,so will choose as the optimal algorithm.This algorithm,project collaborative filtering algorithm on Hadoop platform cannot be implemented because of in the huge database.This article will improve it in the form of a collaborative filtering algorithm based on parallel,and listed as the optimal algorithm,used in this system.So we will improve it as collaborative filtering base on parallel,and apply to the recommendation system.In this paper,the user movie recommendation module is designed from the B / S model,using SSM framework.Firstly,we need to analyze the system environment and the system usage algorithm in this environment,and then analyze the requirements of the system and design the overall framework.The system function module design and database design;finally,the designed system will be tested,each functional module runs normally,reaching the expected recommended effect.