Regresar

Spreading influence values over weighted relationships among users of several social networks

Abstract:

There exist numerous analytic tools for computing the social influence of users. Although it is common to find individuals who actively collaborate in several social networks, existing tools do not consider their overall activity; on the contrary, they compute levels by focusing just on a specific online community. Besides, in most of the existing approaches the users' influence grows with the levels of their contacts, regardless of the kind of relationship established among them. We think that this measure is not fair/objective because the influence of a contact should be just spread to an individual when both take an active part and intensely interact to each other in the social network. This paper proposes a new influence metric which (i) weights dynamically the relationships within a group of individuals by considering their global collaboration in several online communities simultaneously, and (ii) spreads the influence levels over these weighted relationships. As a result, a user's influence increases when he/she is narrowly related to other influential individuals. The benefits of our metric against existing tools have been experimentally validated in a system that promotes free tourist resources and activities through the most influential users in Facebook and Twitter. © 2012 IEEE.