Measuring User Prestige and Interaction Preference on Social Network Site

Measuring user's prestige and interaction preference is an important issue for social network site implementer. To address this problem, an empirical study based on xiaonei.com is conducted. In this work, we introduce the "interaction prestige" metrics based on the observation of user's interaction features. Users with distinct levels of "interaction prestige" are grouped into different categories by a clustering algorithm and named as "outgoing", "reciprocal" and "incoming" type respectively. We demonstrate that the "interaction prestige" of users is related with other prestige metrics including "virtual prestige", "social prestige" and "structural prestige". Further more, the chi-square test for independence shows that there exist significant communication patterns among the identified types of users. We observe that users tend to choose those with high interaction prestige and opposite gender as their communication partner.