Academic journal article Journal of Electronic Commerce Research

Tracing Topic Discussions with the Event-Driven Sir Model for Online Forums

Academic journal article Journal of Electronic Commerce Research

Tracing Topic Discussions with the Event-Driven Sir Model for Online Forums

Article excerpt

(ProQuest: ... denotes formulae omitted.)

1. Introduction

The Internet has supplied new forms of interactivity such as blogs, online forums, and chat rooms [Robbin & Buente 2008]. This interactivity allows users to disseminate information by posting blog entries and discussing their opinions on Web forums. The dissemination process has a considerable influence on society, with numerous citizens and consumers actively expressing their opinions and preferences. Thus, information diffusion in the Web domain has become a major research topic, and studies reveal that it can influence public opinion and society both economically and politically [Flew 2005; Habermas 2006; Robbin & Buente 2008; Woo et al. 2011; Heverin & Zach 2012].

Since the arrival of the social media era, it is necessary for political, economic, and marketing purposes to understand more effectively the mechanism and properties of information diffusion through this new publication method [Wirtz et al. 2013]. For example, firms' strategic decisions can be made by examining a greater number of indepth discussions and their dynamics from the perspectives of participants. Further, firms can leverage the information diffusion process on the Web as a form of viral marketing for their marketing strategies. A careful examination of information diffusion also enables firms to determine how communication with consumers through Web 2.0 channels works and to predict how events such as product launches, promotions, campaigns, and related-news influence consumers.

The most powerful features of this diffusion process are its contagiousness and speed. However, as social media becomes more prevalent [Lai & To 2015], it is easier to propagate not only different views but also misleading or false information. Surprisingly, views found on the Web are sometimes accepted and transmitted to others without much critical examination [Wu & Liu 2008]. This highly contagious property accelerates the information diffusion process. Among the various forms of social media on the Internet, Web forums have been paid special attention as a popular platform for the formation of public opinion, and are considered to be a crucial source for information diffusion whereby users with common interests express and discuss their opinions and influence others. In this regard, interactions occur when an author posts a thread and other authors reply to that thread, thus enabling the author to infect others' opinions.

Many researchers have examined the diffusion mechanism and predicted diffusion behaviors in the fields of epidemiology and sociology. For example, the diffusion process encompasses diseases in biology; computer viruses in networks; and information, knowledge, rumor, social behavior, and innovation in society. Web forums, where participants disseminate as well as receive information online, and form self-contained communities, provide promising opportunities for modeling information diffusion based on the epidemic model. However, despite the attractiveness of Web forums, massive amounts of data and the casual writing styles of participants can make it difficult to analyze opinions on them. This indicates that substantial efforts and advances are needed to collect and analyze highly complex user interactions on Web forums.

This study considers Web forums in order to understand more effectively their information diffusion mechanism and examines whether the formation of opinions about topics on Web forums is similar to the epidemic process. For this, the study considers a large Web forum managed by a major company by using the Susceptible, Infective, and Recovered (SIR) model, a popular epidemic model consisting of mathematical differential equations based on rules of interactions among classes of participants. In addition, the study proposes a novel event-driven SIR model that encompasses the effect of an external event such as news on the number of postings for a given topic. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.