Retweet Predictive Model for Predicting the Popularity of Tweets

摘要

Nowadays, Twitter is one of the most used social networks with over 1.3 billion users. Twitter allows its users to write messages called tweets that now can contain up to 280 characters, having recently increased from 140 characters. Retweeting is Twitter's key mechanism of information propagation. In this paper, we present a study on the importance of different text features in predicting the popularity of a tweet, e.g., number of retweets, as well as the importance of the user's history of retweets. The resulting Retweet Predictive Model takes into account different types of tweets, e.g, tweets with hashtags and URLs, among the used popularity classes. Results show there is a strong relation between specific features, e.g, user's popularity.