Mining and visualizing mobile social network based on bayesian probabilistic model

Social networking has provided powerful new ways to find people, organize groups, and share information. Recently, the potential functionalities of the ubiquitous infrastructure let users form a mobile social network (MSN) which is discriminative against the previous social networks based on the Internet. Since a mobile phone is used in a much wider range of situations and is carried by the user at all times, it easily collects personal information and can be customized to fit the user's preference. In this paper, we presented MSN mining model which estimates the social contexts like closeness and relationship from uncertain phone logs using a Bayesian network. The mining results were then used for recommending callees or representing the state of social relationships. We have implemented the phonebook application that displays the contexts as network or graph style, and have performed a subjectivity test. As a result, we have confirmed that the visualizing of the MSN is useful as an interface for social networking services.