Weakly Supervised Tweet Stance Classification by Relational Bootstrapping

• DOI: 10.18653/v1/d16-1105
• Publication Date: 2016-12-30T08:17:50Z
• Authors: Javid Ebrahimi; Dejing Dou; Daniel Lowd
• Keywords: Bootstrapping (finance)

Supervised stance classification, in such domains as Congressional debates and online forums, has been a topic of interest the past decade.Approaches have evolved from text classification to structured output prediction, including collective sequence labeling.In this work, we investigate stances on Twitter, using hinge-loss Markov random fields (HL-MRFs).Given graph all posts, users, their relationships, constrain predicted post labels latent user correspond with network structure.We focus weakly supervised setting, which only small set hashtags or phrases is labeled.Using our relational approach, are able go beyond stance-indicative patterns harvest more tweets, can also be used train any linear classifier when structure not available costly.