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

With the fast development of network and information technology,massive information emerges rapidly and grows explosively,it is difficult for users to find the necessary information promptly.As an important means of information filtering,the personalized recommendation system is considered to be an effective way to solve the information explosion problem and has been widely used in various fields.Although the personalized recommendation system has made some achievements,but there are still many problems,such as: the quality of recommendation being seriously depressed by data sparseness problems and cold start problems,poor recommendation real-time,poor system expansibility,etc.To solve these problems of current personalized recommender systems,this paper makes a valuable exploratory study on the key techniques of collaborative filtering recommendation algorithm in the personalized news recommendation system,hoping to help readers find the news of interest from the massive news information quickly and accurately.The main work of this paper is:Firstly,summarize the related technologies of personalized recommendation and its research status.Secondly,proposes a new collaborative filtering algorithm based on improved hybrid user model(IHUMCF).First,IHUMCF proposes a new method of user interest modeling,the new interest model based on the traditional user-item rating matrix,and mixed the item attribute and user demographic information,then the user-item rating matrix is transformed into a mixed rating matrix(MFM)based on item attributes and user characteristics,which increases the information density of the user model.Secondly,IHUMCF proposes a new similarity calculation method.In consideration of the different scale of user's rating,IHUMCF introduces the user rating reasonable factor in the user similarity calculation to improve the user similarity algorithm and improve the accuracy of finding the nearest neighbor set.Finally,the paper has designed and implemented the personalized news recommendation system based on IHUMCF,which meets the function of the personalized recommendation,and the performance of the recommendation system has been verified by experiments.Meanwhile,it is also proved that the personalized news recommendation system based on IHUMCF is improved obviously.The proposed IHUMCF recommendation algorithm not only could make up for some deficiencies in the traditional collaborative filtering algorithm,but also could improve the recommendation accuracy,which provides a new idea for the improvement of the personalized news recommendation system.