(A) qualitative method to find influencers using similarity-based approach in the blogosphere

Abstract

Blogs have become an important medium of information in that bloggers convey their own experiences and views on a given topic in a more genuine and conversational manner than can be typically be found in the mainstream media. Accordingly, before making decisions and/or buying brand-new products, internet users may search for information in the blogosphere and read various blog posts. However, to do so they will likely have to read many blog posts as representative influential bloggers are not identified and listed by blog systems. At present, blog systems only list popular bloggers based on traffic information, however, such popular bloggers are not necessarily influential. In the present paper, we present important groundwork for identifying influential bloggers by weighting readers based on their homophily and vulnerability with bloggers. We present the Quantifying Influence Model (QIM), which attempts to measure the influence score of bloggers. QIM is composed of ‘Interpersonal similarity’ and ‘Degree of information propagation’ as follows: (1) ‘Interpersonal similarity’ represents the interaction among bloggers and like-minded readers, and is computed by the tag-similarity. (2) ‘Degree of information propagation’ represents how many readers a blogger has, where the readers diffuse the blog posts via sharing behavior. This quantity is computed as the weighted ratio to infect readers on one’s propensity to share. To evaluate QIM, we conduct experiments with data from Naver blog, which is the largest blog service in Korea and rank bloggers in order of importance. As a result, we find that QIM can differentiate influential bloggers from popular bloggers. More specifically, QIM distinguish readers who have relatively homophilous ties with repeated interactions differently from those who once respond to blog posts simply to utilize resources and those who are non-active. Finally, we compare the frequencies of readers’ referring influential bloggers with th...