From: Characterizing popularity dynamics of hot topics using micro-blogs spatio-temporal data
 | Modeling and predicting popularity dynamics | Spatio-temporal popularity dynamics | Topics popularity dynamics | Our research |
---|---|---|---|---|
Model and algorithm | Rank-shift model [1] SEISMIC algorithm [16] Branching process [26] | K-SC clustering algorithm [30] Popularity growth model [2] | Reinforced poisson processes [28] Support vector regression [8] | K-SC clustering algorithm |
Data source and form | Wikipedia topics and an Web pages [1] 3.2 billion tweets and retweets on Twitter [16] | A set of 580 million Tweets, and a set of 170 million blog posts and news media articles.[30] Meme [24] | 196 million tweets from 10 million users.[42] Hashtags in twitter [5] | 1259 hot topics between October 4, 2013 and November 4, 2013. 138,609 microblogs related to these topics |
Methods and tools | Simulation by Matlab [1] Statistics by R [16] | Clustering by Matlab [30] Spike trains [17] | Statistics by Excel [42] | Statistics by Excel Simulation by Matlab |
Main findings | Dynamics of popularity are characterized by bursts [1] only 15% relative error in predicting [16] | Can predict the shape of volume over time with the accuracy of 81% [30] | Highly popular topics are those that cross regional boundaries [42] | Topics popularity is subject to the power law form |