De retibus socialibus et legibus momenti

TL;DR

This paper compares existing influence measurement techniques in online social networks using click data and introduces a novel, physically-inspired method for assessing influence.

Contribution

It presents a new influence measurement approach based on a physical metaphor and evaluates it against existing techniques using click data in OSNs.

Findings

The new method shows comparable or improved correlation with user influence.

Existing techniques vary significantly in effectiveness.

Physical metaphor provides an intuitive framework for influence assessment.

Abstract

Online Social Networks (OSNs) are a cutting edge topic. Almost everybody --users, marketers, brands, companies, and researchers-- is approaching OSNs to better understand them and take advantage of their benefits. Maybe one of the key concepts underlying OSNs is that of influence which is highly related, although not entirely identical, to those of popularity and centrality. Influence is, according to Merriam-Webster, "the capacity of causing an effect in indirect or intangible ways". Hence, in the context of OSNs, it has been proposed to analyze the clicks received by promoted URLs in order to check for any positive correlation between the number of visits and different "influence" scores. Such an evaluation methodology is used in this paper to compare a number of those techniques with a new method firstly described here. That new method is a simple and rather elegant solution which tackles with influence in OSNs by applying a physical metaphor.