Networked individuals predict a community wide outcome from their local information

Chesney, Thomas

Abstract

The term ‘viral’ is used to describe a phenomenon that tends to be shared by those who encounter it. This paper considers the act of responding positively to a phenomenon by sharing it with others, something exemplified by the online social media acts of choosing to ‘like’ on Facebook, ‘retweet’ on Twitter, or by a similar mechanism on websites such as LinkedIn, Flickr or Pinterest. Using a threshold model of influence, simulations are run on four network structures where a critical mass chooses to share a phenomenon that eventually either goes viral or does not. The data collected are examined to determine whether an individual node can make an accurate prediction about the state of the entire network just from information on the behavior of their neighbors. The intention is to study what it is in terms of network structure that makes an individual good at sensing the zeitgeist, or ‘spirit of the age’.

Findings show that those best placed to predict are ‘important’ as measured by network centrality, and members of numerous communities. The characteristics of the critical mass are important in determining the spread of a phenomenon and it is possible for an individual node to predict an outcome as well as an observer who has access to the state of every node in the network.

Potential applications might be found in monitoring the success of marketing campaigns, or in organizations wishing to keep abreast of current trends in a situation where data on network structure is available but data on the activity of network members is limited.

Citation

Chesney, T. (2014). Networked individuals predict a community wide outcome from their local information. Decision Support Systems, 57, 11-21. https://doi.org/10.1016/j.dss.2013.07.006

• Journal Article Type: Article
• Acceptance Date: Jul 24, 2013
• Online Publication Date: Jul 31, 2013
• Publication Date: 2014-01
• Deposit Date: Feb 21, 2018
• Journal: Decision Support Systems
• Print ISSN: 0167-9236
• Electronic ISSN: 1873-5797
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 57
• Pages: 11-21
• DOI: https://doi.org/10.1016/j.dss.2013.07.006
• Keywords: Social network; Consensus decision; Viral marketing; Simulation
• Public URL: http://www.sciencedirect.com/science/article/pii/S0167923613001917#
• Publisher URL: https://www.sciencedirect.com/science/article/pii/S0167923613001917?via%3Dihub