Common neighbour similarity-based approach to support intimacy measurement in social networks

A large amount of social data is being generated every day, as the Internet becomes more pervasive and mobile devices more ubiquitous. Accordingly, Internet users often experience difficulty finding the content they want, resulting in the popularity of personalized services that provide user-customized content. Intimacy between users of social network services can be utilized as a foundational technology for personalized services. In this paper, an intimacy measurement method for social networking services based on common neighbour similarity is proposed. The proposed method uses the link relationship between users for intimacy measurements and can be applied to general users. Further, it promotes easy data collection using publicly available data. To evaluate the proposed intimacy measurement method experimentally, a significant amount of user data was collected from Twitter. In addition, various statistical datasets were presented, and regression analyses conducted on graphs extracted from user data were collected to interpret the meaning of the intimacy index measured using the proposed method with existing social networking services.