Modeling the structure and evolution of discussion cascades

Abstract

We analyze the structure and evolution of discussion cascades in four popular websites: Slashdot, Barrapunto, Meneame and Wikipedia. Despite the big heterogeneities between these sites, a preferential attachment (PA) model with bias to the root can capture the temporal evolution of the observed trees and many of their statistical properties, namely, probability distributions of the branching factors (degrees), subtree sizes and certain correlations. The parameters of the model are learned efficiently using a novel maximum likelihood estimation scheme for PA and provide a figurative interpretation about the communication habits and the resulting discussion cascades on the four different websites.

Generative models of online discussion threads: state of the art and research challenges

“…This discrete-time model extends previous generative models for discussion threads [83,92]. Besides popularity and novelty features (parameterized with α and τ respectively), it considers an additional feature (a root bias) that makes explicit the difference between the process of writing to the post (with id 0) node and to a descendant (user comment).…”

“…While the threads sizes analyzed in some studies followed heavy-tailed distributions [83,92], other previous studies reported distributions with a light tail [93,94]. Wang et al [84] first observes that the waiting time between two consecutive comments from a user follows a upper truncated Pareto distribution.…”

Handbook of Technical Communication

“…Any account of technical communication has to consider this social dimension (cf. Zerubavel, 2003) -otherwise it fails to clarify the peculiarities of technological enhancements of, say, collaborative writing as currently provided by wiki-based systems (Leuf and Cunningham, 2001;Stegbauer, 2009 (Wattenberg, Viégas, and Hollenbach, 2007;Laniado, Tasso, Volkovich, and Kaltenbrunner, 2011;Gómez, Kappen, and Kaltenbrunner, 2011) (Mehler and Sutter, 2008)) to collaborative writing and edit wars (where articles are authored by many authors in the role of content providers, removers or defenders (Brandes, Kenis, Lerner, and Raaij, 2009) (Sager, 1997). They also allow for suspending the time-space continuity of faceto-face communication (Zerubavel, 2003) in terms of multilogues.…”

Social contagion with degree-dependent thresholds

“…As an early example, Granovetter suggested that "the number or proportion of others who must make one decision before a given actor does so" is a fundamental personal trait driving the contagion [2]. It is straightforward to see that this type of dynamics is sensitive to network structure [5][6][7][8][9][10]. This paper investigates what happens if a threshold, such as the one Granovetter mentions, depends on the how influential a node is.…”

“…Since Granovetter, there has been many papers investigating the relation between the network structure and opinion dynamics [5][6][7][8][9][10]. Perhaps the most influential paper was Watts's study of threshold models as an explanation of cascades or fads [5].…”