Study On Social Network Friends Recommendation Method Based On Trust Relationship And Interest Model

With the explosive growth of users in the social network,users share massive information everyday,which causes information overload,therefore,it makes users more difficult to seek interesting information.An effective solution is to recommend relevant and high quality sources of information, the recommendation of friends,which can not only help users to obtain useful information,but also meet user's needs to communicate with users who have the same similar interests.However, the large amount of users and their diverse and dynamic interests possess great challenges to recommend potential friends in social network.This thesis focuses on the trust relationship and user's interest and preference for the potential friend recommendation in social network.By analyzing the characteristics of relationship of users in social network,this thesis proposes social network friends recommendation method based on trust relationship.And then by analyzing the characteristics of text information of users in social network,this thesis builds user interest model,and then combines the trust relationship and user interest model to propose potential friends recommendation method based on trust relationship and user interest model.The content of this thesis is mainly as follows:(1)Analysis of the advantages and disadvantages of the recommendation methods and the characteristics of social network, to provide the theoretical basis for the social network potential friends recommendation method.(2)Quantify the user relationship information and micro-blog contents implicit information, to provide the basis for further data mining in social networks.(3)Use the social relationship between users and the micro-blog content to build trust network and user interest model.(4)Propose friends recommendation method based on trust relationship and user interest model.(5)Design and implement a potential friends recommendation system based on trust relationship and user interest model and verify the effectiveness of the friends recommendation method.Finally,Using the Micro-blog real-world data sets,experimental results and comparisons with the traditional recommendation method show that the proposed recommendation method based on trust relationship and interest model outperforms traditional recommendation method and social network friends recommendation method based on trust relationship.The proposed method can effectively improve the performance of potential friends recommendation.