Influence diffusion detection using blogger's influence style

Abstract

Previous studies on detecting blogosphere influence diffusion had used blog features such as in-degree and sentiment links. The approaches in most of these studies assumed that influence increases with the number of links and largely ignored the possible effect of bloggers' influence style on the diffusion of influence between linked bloggers where influence could be further described through the engagement style, persuasion style, and the persona of the bloggers. In this paper, we propose an Influence Diffusion Detection Model - Influence Style (IDDM-IS) that includes the use of bloggers' influence styles to detect influence diffusion through the blogosphere. Our study analyzed 107 bloggers with varying influence styles to detect the influence diffusion path. The results showed performance for IDDM-IS to be better than the in-degree and sentiment-values baseline approaches. In addition, IDDM-IS could provide a fine-grained description of the influence diffusion paths using the bloggers' influence styles. © Springer International Publishing 2013.