User Character and Communication Pattern Detecting on Social Network Site

As a newly emergent interpersonal communication media, social network site has been attracting researchers\´ attention world wide. Aiming at disclosing user characters and interpersonal communication pattern, an empirical study based on xiaonei.com is conducted. User\´s profile data of a typical university has been crawled based on which four types of interaction are identified. Then, these types are used as each user\´s feature vector and taken as the input of a clustering algorithm. Three categories of participant are found which are named as "outgoing", "incoming" and "reciprocal" users. Furthermore, the chi-square test for independence conducted on the interaction data shows that there exits significant communication patterns among different categories of users. This computation method can facilitate user data analysis and user interface design for Social Network Site implementers.