Viral Marketing in Social Networks
Our recently-accepted paper A Scalable Heuristic for Viral Marketing Under the Tipping Model (to appear in Springer’s Social Network Analysis and Mining, by MAJ Paulo Shakarian, LT Sean Eyre, and CDT Damon Paulo) was featured in MIT Technology Review this week. In this paper, we consider the tipping model on a social network, where an individual adopts a new behavior, product, or other attribute based on a percentage of his friends having previously made the same choice. This leads to a cascading effect that has previously been observed in other experimental studies. Our challenge was to find a small set of individuals such that if they initially adopt the behavior, it spreads throughout the entire social network. These are called “seed sets.” We developed an algoirthm that does so and was able qucikly identify very small seeds sets (often less than a percent of the population) in extermely large social networks (millions of nodes and edges). Previously, I gave a talk on this topic, you can view the video of it on YouTube. In the paper, we present a series of results that study the behavior of the algorithm in a variety of settings, including how it performs when hgih-degree nodes are not available for tageting, how the cascade process proceeds over time, and what characteristics of a network make it more susceptible to viral marketing. A pre-print of the paper is available from arXiv. For more information on this paper, please contact MAJ Paulo Shakarian. This work was funded by the Army Research Office.