High Correlation between Incoming and Outgoing Activity: A Distinctive Property of Online Social Networks?

Abstract

User influence is an important topic of research for online social networks. Recent work has shown that a user's influence is not directly related with node in-degree. However, the definition of what is an influential or relevant user is still an open subject. In general, we can say that an influential user has the ability to produce incoming activity in an interaction graph. For this reason we have focused our attention on the user's incoming activity and the search for which factors are related with this indicator. We have studied a Facebook dataset and have found that outgoing activity is highly correlated with incoming activity. This characteristic is valid not only for users with a low level of activity, but it is also valid for high activity users. This result hasn't been reported before and appears to be an important factor of user behavior. To contrast this finding, we have compared it against a popular e-mail dataset and a Twitter dataset. The result was that we found a similar behavior for the Twitter Dataset, but for the e-mail dataset the correlation was lower. Hence, we conjecture that the high correlation may be a distinctive property of social networks. In our future work, we propose to extend this study to other social network platforms.