Research On The Estimation Method Of Social Network Relationship Strength

Social Network Service has become an indispensable part in our daily life and played a more and more important role in people's communication. People in the online social networks would link with their friends through chatting, e-mail, posting, commenting directly, that means these behaviors could intuitively and effectively map the users' social situation in reality to the social network. It can be said that the real social relationship is the foundation of social network relationship, and the social network platform is the carrier of the further development of social relationship.However, the present link-ship in Facebook, Myspace or Sina Weibo, provides only a coarse representation of relationship like “Follow”, “fans”, “friends” other than revealing the detail closeness. This simple binary status may confuse the acquaintance and the close friend without their relationship strength. Actually, people obviously prefer contact with their friends to just acquaintances and more inclined to exchange emotion with their close friends. Therefore, it's more reliable to use relationship strength to judge the closeness than just simple relationship status. Moreover, it is found that the weak relationship plays a more important role in non-repetitive information dissemination, the opinion spreading and the satisfactory job searching.According to the importance of relationship strength calculation in social network, this paper will estimate the relationship strength from two aspects. On one hand, we propose a probabilistic graphical model to measure the relationship strength between different users by taking consideration of the similarity of users' profiles, co-occurrence of user names and interaction activities in different activity fields, the users here have already be friends. On the other hand, in the light of non-friendship users, we estimate their latent relationship strength based on their inherit relationship strength and their propagate relationship strength with the profile similarity, social document similarity, social friend coincidence and the friendship strength obtained from last step. Lastly, according to the ranking of the latent relationship strength, the recommending of Top K potential friends completed. In view of the above two relationship strength estimation methods, this paper reveals the evaluation result compared with the existing research work, and verifies the rationality of our paper.