Predictability of extreme events in social media

Abstract

It is part of our daily social-media experience that seemingly ordinary items (videos, news, publications, etc.) unexpectedly gain an enormous amount of attention. Here we investigate how unexpected these events are. We propose a method that, given some information on the items, quantifies the predictability of events, i.e., the potential of identifying in advance the most successful items defined as the upper bound for the quality of any prediction based on the same information. Applying this method to different data, ranging from views in YouTube videos to posts in Usenet discussion groups, we invariantly find that the predictability increases for the most extreme events. This indicates that, despite the inherently stochastic collective dynamics of users, efficient prediction is possible for the most extreme events.

13 pages, 3 figures

References in corpus (8)

• Statistical physics of social dynamics
• Robust dynamic classes revealed by measuring the response function of a social system
• Novelty and Collective Attention
• The Matthew effect in empirical data
• The Pulse of News in Social Media: Forecasting Popularity
• Niche as a determinant of word fate in online groups
• A maximum entropy framework for non-exponential distributions
• Precursors of extreme increments

Cited by in corpus (4)

• A maximum entropy framework for non-exponential distributions
• Twitter-based analysis of the dynamics of collective attention to political parties
• Stochastic dynamics and the predictability of big hits in online videos
• Predicting Engagement with the Internet Research Agency's Facebook and Instagram Campaigns around the 2016 U.S. Presidential Election