Study On Application Of Recommendation Algorithm Based On Social Network

With matures and the rapid development of Internet Technology, social networks have went deeply into every corner in People’s Daily life. Based on the users’ browsing history data and friends relationship, the recommendation system analysis the users’ and friends’ hobbies, implicit and actively referring items to people, has gradually become one of the important ways for people to obtain information on the social network. How to introduce items to people in his interesting effectively, and obtaining the user’s approval, to obtain a higher breakthrough rate, is the core issue of recommendation algorithm studies.Firstly, this paper introduces the research status about recommendation system and social networks at the present, and analyzes the application scope and the advantages and disadvantages of different recommendation algorithm involved personalized recommendation system. Then combining with the structural features of social network, we analyze influence on the personalized recommendation system of the social network relations, information dissemination way and the time factor’s. The current recommendation algorithm of social network making inadequate use of personalized information of friends and the item, also not fully embodies personalized relationship features in social network between user and items. There is a large amount of calculation and high complexity in the algorithm process. Therefore, this paper focuses on how we can making full use of social networking features to improve personal recommendation algorithm.Secondly, the paper analyzes similarity relations between the personalized recommendation in the social network and figure ranking problems, and the Page Rank algorithm is introduced which is based on graph ranking problem, and also introduce Personal Rank personalized recommendation algorithm based on random walk thought Page Rank. What’s more, according to the structure characteristics of the social network, we propose the concept and calculation method of the relationship strength between users and the users’ intensity in the items. Then, it discusses the significance both in social networking recommendation algorithm, and the concept of weights will be unified join in the calculation formula of the Personal Rank recommendation algorithm, so as to achieve the efficiency of the improved algorithm. Finally the correctness and validity of the algorithm is verified by experiment.