Design And Implementation Of A Personalized News Recommendation System

As the concept of "Internet +" becomes an indicator of the Internet and traditional industries of collaborative innovation,news media on the Internet are at the forefront of innovation.Under the background of mass news,how to transfer mass news to the most needed readers,as well as a more comprehensive and accurate understanding of the needs of the reader's interest,tailored personalized news system has become urgent attention,and the urgent problem should be solved.In this context,the special features of news are analyzed,and key technologies and algorithms used in news recommendation,a news recommendation model based on collaborative filtering algorithm and content-based thinking is designed,to provide readers with personalized news reading experience.The main work in the paper is summarized as follows:Firstly,the key technologies and algorithms of the personalized news recommendation are introduced,and paper focus on the commonly used recommendation technologies and algorithms,such as content-based and collaborative filtering which including the recommendation algorithms based on item,user and slope-one.Secondly,a news recommendation model based on collaborative filtering algorithm is designed,which filtering the news in a way that is based on the news content,and remove duplicate,and the collaborative filtering is the core of personalized news recommendation.Finally,increasing the heat news and local news in the news recommendation results,as a result of CF supplements,and enriched personalized news recommended content.Thirdly,collecting the news and reader information as the original data to achieve the personalized news recommendation system,and apply it to a news media recommendation project.Through the test data set to test,the analysis of precision and recall of the results shows that the system can achieve good results of personalized news recommendation.