Joint Behavior-Topic Model for Microblogs
We propose an LDA-based behavior-topic model (B-LDA) which jointly models user topic interests and behavioral patterns. We focus the study of the model on on-line social network settings such as microblogs like Twitter where the textual content is relatively short but user inter-actions on them are rich.
Reference: It's Not What We Say But How We Say Them: LDA-based Behavior-Topic Model, Minghui Qiu, Feida Zhu and Jing Jiang, 2013 SIAM International Conference on Data Mining (SDM'13), Austin, Texas, USA, May, 2013.
Contact: minghui.qiu.2010@smu.edu.sg