Multidimensional community detection in Twitter

Abstract

We present and apply a generic methodology for multidimensional community detection from Twitter data. The approach builds on constructing multiple network structures based on the similarity and interaction patterns that exist between different users. It then applies traditional network centric community detection techniques to identify clusters of users. The paper also approaches the issues of dynamicity and evolution in Social Media by developing a Bayesian classifier that maps new users to the detected communities. Using a data set of UK political Tweets, we evaluate the factors affecting the quality of the detected communities. We also investigate how the accuracy of the classifier is affected by the dynamicity of the network evolution and the time elapsed between community detection and classifier application.