Research On Privacy Protection Of Social Network Recommendation Systems

With the development of information technology and Internet, network life gradually become social. The amount of data in the network grows rapidly, how to find the interested information is becoming more and more important in the huge network data. Personalized recommendation system can solve these problems, and it can not only improve customer loyalty, satisfaction, but also can improve the benefits of the companies.In addition to the personalized recommendation system, security and privacy of individual users has also become an important privacy control social network development. A good recommendation system prove highly personalized, accurate and efficient recommendation, it is necessary to fully understand and grasp the user's personalized information and requirements. At the time of collecting the information,we may didn't pass the user's permission sometime.As the ego to protect consciousness enhancement, the user is becoming more and more concerned about their privacy. When a user find their search history been stolen and used by others,they become dislike and distrust of personalized recommendation. Therefore, to solve safety problems in social networking recommendation system, to promote development in the recommendation system.Moreover, in order to solve the problem of user privacy leakage in social networking recommendation system, this article we put anonymous privacy into the recommended links, and improve the conventional recommendation algorithm.Put forward a "Collaborative filtering hybridrecommendation system based on node anatomyprivacy protection"(hereinafter referred to as NAPPHRS). It can not only effectively protect user information, personalized recommendation and can efficiently and accurately.In this paper, the main research work and achievements are as follows:(1)Personalized recommendation technology are reviewed in this paper, the privacy protection technology and the recommendation system of privacy protection technology research present situation, points out that the recommended system needs further and research problems, namely the user privacy issues.(2)The principle of the proposed recommendation system and the implementation process has carried on the detailed elaboration, mainly includes threemain: based on the recommendations from the content, rule-based recommendation and collaborative filtering recommendation; Propose the recommendation system based on user clustering and project properties algorithm combining constructed combination recommendation algorithm.(3)In the related concepts of social network and social network security issues,including personalized recommendation system and the network users' privacy security connection and influence. For later attribute to privacy protection algorithm based on node split.(4)For sorting process, personalized search results on the server side using user description file to user privacy problem, Add the privacy protection based on node split to the combination collaborative filtering algorithm,comes out of the design of NAPPHRS. Through the simulation experiment to validate its feasibility and validity,the experimental results show that the proposed system can to protect user information and attributes, with a better accuracy and coverage result.In this paper, the main innovation points are as follows:(1) Considering the social network users correlation property distribution, the nodes segmentation algorithmwas proposed based on attribute partition. Segmentation process keep the correlation characteristics of property distributionas much as possible, result in improving the anonymity of the users with privacy attribute node.(2) Put the privacy protection algorithm based on node split into the process of personalized recommendation, improve the privacy protection of personalized recommendation, well realized the original intention of this article.(3) Hybrid collaborative filteringrecommendationreducedthe impact of weak data availability, result withan ideal output.