Research And Implemention Of Social Network Information Recommendation Methods That Support Privacy Protection

With the development of information technology and the popularization of Internet,social network has gradually become an important part of people's daily communication.However,due to the rapid growth of information,people enjoy the convenient services brought by social media,at the same time,how to find the content they are interested in in in the massive data has become a problem.In this context,the social network recommendation system came into being,at the same time,the social network recommendation system also has the risk of user privacy disclosure.How to ensure the privacy security of users and provide high-quality recommendation service has become a research hotspot.At present information recommendation mainly focuses on advertisement recommendation,location-based recommendation and e-commerce network recommendation.However,there are few researches on social network information recommendation supporting privacy protection.This paper mainly studies how to recommend interested blogs to users while protecting their privacy,proposes the methods of guessing users' interests and recommending social network information to support privacy protection,and uses Sina Weibo public data set for verification.Experiments show that the accuracy and recall rate of user interest estimation method is more than 80%,and the privacy loss ratio of privacy protection method is less than 20%.The main contributions of this paper are as follows1.User interest prediction based on social network is proposed in this paperTo solve the problem that it is difficult for social network users to get the information they want from the Internet,this paper proposes a method of social network user interest speculation.Through the classification features of friends'publishing,commenting and forwarding blog posts,we get friends' interests,and use personalrank algorithm to calculate friends' interests,to solve the cold start problem of user rating data;through friends' interests,familiarity between users and friends,and interest similarity,we calculate users' interests,and take the first n interests as the inferred user interests.The experimental results show that the proposed method is effective in inferring user interest.2.A privacy protection method supporting information recommendation is proposedIn view of the problem of privacy disclosure caused by information dissemination in social network information recommendation,a method of social network information recommendation supporting privacy protection is proposed based on information dissemination model.Users are allowed to set Limited access user list through personalized preference settings,and the probability of information dissemination to blacklist users is calculated by using the list of limited access users;the threshold value of privacy disclosure is set to describe the probability boundary value of access to privacy blog by blacklist users,so as to protect users' privacy in information recommendation.The experimental results show that the proposed method can protect the privacy of users while ensuring the recommendation effect.3.The prototype of information recommendation system supporting privacy protection is proposedIn view of the risk of disclosing the user's privacy of the recommended information in the recommendation system,a prototype of information recommendation system supporting privacy protection is developed,which realizes the functions of user interest speculation,privacy protection,information recommendation and data visualization.Through the data visualization function,the user interest and user influence are displayed,in which the user interest font shows the user interest,the user bubble shows the user influence,and the visualization method is used to display the data directly,which is conducive to the further analysis of the data.