Design And Implementation Of Personalized News Recommendation System Based On Collaborative Filtering Algorithm

With the continuous expansion of information,processing information has become an increasingly difficult task for peopl e since it is hard to filter and obtain information efficiently.As a result,people fail to make quick decisions when confronted with the flow of information.The recommendation system can effectively process and filter information,delivering effective information to users,thus it can greatly shorten time for users to get it.News system provides users with access to news.In today's society where the speed of producing news is getting faster,people find it harder to obtain the news they are interested in.In addition,due to the wide spread popularity of new media and ‘we media',the quality and feasibility of news have been threatened,it is more and more difficult for people to filter news information.Therefore,by introducing recommendation system to news system to filter and deal with information can greatly improve users experience,as well as helping users to get rid of fatigue when reading news.Personalized news recommendation system can analyze users interests according to their browsing history,further exploring users' interests via users tags and thus recommending news information they care about.Base on this context,this paper adopts recommendation algorithm to recommend news to users.The main contents of this paper are as follows:First of all,it helps to direct research by analyzing the source and purpose of personalized news recommendation system as well as the present domestic and overseas development condition in detail.Then based on this,this paper further analyzes related technologies adopted in this system,including recommendation system and recommendation technology.Secondly,detailed requirement analysis and design for news recommendation system are carried out.The requirement analysis has two aspects,the function requirement and performance requirement.Then the design is mainly concern about the overall systemetic structure,process analysis and functional module design.Among them,the overall systemetic structure is further analyzed from two aspects,systemetic structure and network deployment.System function modules can be divided into four parts,including user management module,theme management module,personalized recommendation module and news display module.The four modules were designed specifically.Finally,the personalized news recommendation system is fully implemented and tested.This system mainly uses Java as the development language and Eclipse as the development platform.The database uses Mysql and the file storage uses HDFS.The user management module,the theme management module and the personalized recommendation module are implemented and realized.The user management module is implemented from adding and deleting data,etc.The theme management module uses Chinese word segmentation and frequent mining to discover theme models.Thhis paper adopts Jcseg word breaker and Apriori frequent set mining algorithm.Personalized recommendation module achieve s to realize similarity calculation module of users or news and personalized recommendation sub-module.