Research For Friend Recommendation Algorithm In Online Social Networks

With the rapid development and popularity of online social networks, the data in social networks increases with an exponential growth trend. Due to the large amount of information, users in social networks cannot deal with it effectively. The useful content in the social network data is not increased, but showing a decreasing trend. The search engine, just identify the object to be served based on the user’s query without considering the user’s personality. This cannot lead to effectively satisfied query results. Information explosive growth leads to certain challenge for search engine, but also makes it difficult for users to expand their social circle. It cannot have good user experience in social networks. Also it causes users to lose their stickiness of social networks. All of these will harm businesses profit. The users recommendation as an important content in the social network, by recommending the right friends to users, can efficiently help users find useful information. Hence, how to accurately and efficiently recommend friends to the users will be a challenge and useful research problem in the future.(1) In order to reflect the tendency of users in the recommendation of friends as long as the degree of interaction relationship between people in real life, this paper considers the interaction between the users by taking into account the direction of the interaction. A trust degree is defined based on the friend relationship between the users.(2) Based on the trust degree, a social circle detection algorithm is proposed. Firstly, the algorithm improves the similarity measure by incorporating the topological information of the adjacent edges in the social graph via fusing the trust relationship among the users. Then the content information released by users is combined. A new method to calculate the similarity between adjacent edges is proposed. The user’s trust social circle is derived based on this new algorithm.(3) Based on the trust social circle, a friend recommendation algorithm is proposed. The proposed algorithm improves the method of the similarity between users by considering the trust degree between users in the social circle, so as to achieve the user’s friends recommendation.(4) The experiment is conducted to compare the proposed algorithm with the current advanced friend recommendation algorithm and the traditional friend recommendation algorithm using facebook data.The experimented results verify the accuracy and effectiveness of the proposed algorithm to a certain extent.