RT to Win! Predicting Message Propagation in Twitter

By Sasa Petrovic, Miles Osborne, Victor Lavrenko. International Conference on Web and Social Media, 2011

430 citations recorded by the source corpus.

Twitter is a very popular way for people to share information on a bewildering multitude of topics. Tweets are propagated using a variety of channels: by following users or lists, by searching or by retweeting. Of these vectors, retweeting is arguably the most effective, as it can potentially reach the most people, given its viral nature. A key task is predicting if a tweet will be retweeted, and solving this problem furthers our understanding of message propagation within large user communities. We carry out a human experiment on the task of deciding whether a tweet will be retweeted which shows that the task is possible, as human performance levels are much above chance. Using a machine learning approach based on the passive-aggressive algorithm, we are able to automatically predict retweets as well as humans. Analyzing the learned model, we find that performance is dominated by social features, but that tweet features add a substantial boost.

Subfields: Sociology and Political Science, Statistical and Nonlinear Physics

View Paper: https://www.semanticscholar.org/paper/6952cc377adecd16e4f09b15c56d83aa8c86878d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Sociology and Political Science

In Statistical and Nonlinear Physics

Citations in the corpus

Can cascades be predicted? (2014, methodology): “At the other end of the spectrum, research has focused on individual userlevel prediction tasks: whether a user will retweet a specific tweet [26] or share a specific URL [10].”
DeepHawkes: Bridging the Gap between Prediction and Understanding of Information Cascades (2017, background): “These features mainly include temporal features [2, 3], structural features [16, 18, 19], content features [15, 20–22] and features defined on early adopters [13, 23, 24].”
Context-aware Image Tweet Modelling and Recommendation (2016, background): “We first describe the four strategies that construct contexts from different data sources (cf. Figure 2), and then discuss the rules to fuse the contexts which help to improve text quality and save the acquisition cost.”
SEISMIC: A Self-Exciting Point Process Model for Predicting Tweet Popularity (2015, neutral): “Keywords: information diffusion; cascade prediction; self-exciting point process; contagion; social media.”
The Structural Virality of Online Diffusion (2015, background): “Whether or not such cases appear with any meaningful frequency is, however, largely an empirical matter, and hence the utility of the metric must ultimately be evaluated in the context of real examples, which we discuss further below.”
What makes an image popular? (2014, background): “While most of the work has focused on predicting popularity of text content, such as messages or tweets on Twitter [42, 20], and some recent works on video popularity [43, 49, 38], signiﬁ-cantly less eﬀort has been expended in prediction of image popularity.”
A Proactive Intelligent Decision Support System for Predicting the Popularity of Online News (2015, methodology): “Similarly to what is executed in [12,3,7], we assume a binary classification task, where an article is considered “popular” if the number of shares is higher than a fixed decision threshold (D1), else it is considered “unpopular”.”
Predicting Responses to Microblog Posts (2012, neutral): “(2010) and Petrovic et al. (2011) predicted retweets.”