Abstract

Twitter's list feature allows users to organize their followees into groups for easier information access and filtering. However, the percentage of users using lists is very small and most existing lists have only a few members. One reason for this may be that curating groups of Twitter users is a time consuming task. In this paper, we propose early and late fusion methods for automatically clustering followees using both graph structure and tweet content. We evaluate our approaches using ground-truth Twitter lists crawled via the Twitter API and show that the late fusion method outperforms both the baselines and the early fusion method.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call