Twitter is a strange beast. Sometimes there is no telling when the Twittersphere will latch onto something and make it go viral but some researchers have got a model able to predict the likelihood of how far and how fast a given tweet will be retweeted.
The model was created by Tauhid Zaman (MIT assistant professor), Emily Fox (University of Washington) and Eric Bradlow (University of Pennsylvania) and was presented in a paper to the Annals of Applied Statistics. The group examined retweets on a variety of topics, looking at how fast these retweets were spread. The model itself relies on how quickly a new tweet is picked up and retweeted in order to come to its conclusions, based on the behaviour of the retweets examined.
The retweet prediction model is obviously not spot-on but it does have applications when it comes to “understanding retweet behavior [which]could lead to a better understanding of how broader ideas spread in Twitter and in other social networks.”
Source: Technology Review