Research On News Recommendation System Based On Community Network

The advent of the era of big data has brought convenience to the sharing and dissemination of information,but the vast amount of information has also caused the problem of information overload.The emergence of recommendation system has become one of the effective means to solve the problem of information overload.News recommendation is one of the important research fields.News recommendation system obtains user's reading interest by analyzing user's news browsing history information,and uses the recommendation algorithm to recommend news information that meets user's interests and preferences,so as to solve the problem that users are difficult to obtain information of interest.In the common news recommendation algorithms,collaborative filtering algorithm has some problems,such as rough similarity calculation and high computational cost.Content-based recommendation algorithm can not reflect the change of user interest in time.This paper studies the problems existing in the above methods.In this paper,based on the problems existing in the collaborative filtering algorithm,the idea of network community is introduced into recommendation algorithm,and a recommendation algorithm based on community detection is proposed.Firstly,the similarity of reading behavior and content similarity between users are calculated to form the user's comprehensive similarity,and the user relationship network is constructed.Then the community is divided according to the Fast Unfolding algorithm,and the users with similar interests are divided into the same community.In the community,collaborative filtering algorithm based on user comprehensive similarity is used for recommendation.Whensearching for neighbor user set,it is no longer to calculate all user data,but to target users in the community where the target user is located,narrowing the range of calculation of neighboring users.According to the problems of content-based recommendation algorithm,this paper introduces a time-decay function to calculate the change of user's interest in different time based on content-based recommendation algorithm,and the user long-term interest model is established.The community discovery algorithm is used to obtain the neighbor user of the target user,and the user potential interest model is established according to the interest model of the neighbor set user,and then the two user interest models are mixed to obtain the user mixed interest model for use in the recommendation algorithm for news recommendation.Finally,the experiment data set provided by DataCastle was used for experiments.The experimental results show that the proposed algorithm has improved the accuracy,recall rate and F-measure value compared with other algorithms,which improves the accuracy of recommendation and meets the user's personalized needs.