Comparing Social Influence with Collaborative Influence in Human Flesh Search

Previous studies believe that the individuals who have great social influence play an important role in the collaborative activity. However, in Human Flesh Search (HFS), which is a typical collaborative activity originated from Chinese online social networks, we do not obtain the same conclusion but find that the correlation between social influence and collaborative influence is weak. Specifically, we first construct two networks from the crawling data, i.e., the social network and collaborative network of HFS participants, and quantify social influence and collaborative influence separately. Then we calculate the correlation between social influence and collaborative influence. Furthermore, we obtain three different types of participants by using BIC measure and k-means++ clustering algorithm and calculate the correlation between social influence and collaborative influence among different types. Although there exist differences in different types, all the results show that the correlation between social influence and collaborative influence is weak.