Discovering and Predicting Influential Users During Urgent Diffusion Events on Social Media

• Activity Number: 552
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Hechao Sun* and Shawn Mankad and William Rand
• Companies: University of Maryland Robert H. Smith School of Business and University of Maryland and University of Maryland
• Keywords: urgent information diffusion ; retweet cascades ; influential nodes ; cascade size prediction

Abstract

Given information up to time T, we manage to find metrics to forecast a group of users who are likely to be influential in spreading information during urgent events on Twitter during some future time after T. We use the size of retweet cascades generated by the given user as the empirical measure of influence. For those selected nodes, we also try to make prediction about their cascade size, where we mainly adopt their network features of retweet network up to time T. We have shown that the dynamic network features, which we have not seen in previous related work as far as we are concerned, can much improve the prediction. Finally, we suggest reasonable communication strategies based on our findings for managers involved in critical events.