Leveraging Behavioral and Social Information for Weakly Supervised Collective Classification of Political Discourse on Twitter

Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, pages 741–752 Vancouver, Canada, July 30 - August 4, 2017. c⃝2017 Association for Computational Linguistics https://doi.org/10.18653/v1/P17-1069 Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics, pages 741–752 Vancouver, Canada, July 30 - August 4, 2017. c⃝2017 Association for Computational Linguistics https://doi.org/10.18653/v1/P17-1069 Leveraging Behavioral and Social Information for Weakly Supervised Collective Classification of Political Discourse on Twitter Kristen Johnson, Di Jin, Dan Goldwasser Department of Computer Science Purdue University, West Lafayette, IN 47907 {john1187, jind, dgoldwas}@purdue.edu Abstract Framing is a political strategy in which politicians carefully word their statements in order to control public perception of issues. Previous works exploring polit- ical framing typically analyze frame us- age in longer texts, such as congres- sional speeches. We present a collec- tion of weakly supervised models which harness collective classification to predict the frames used in political discourse on the microblogging platform, Twitter. Our global probabilistic models show that by combining both lexical features of tweets and network-based behavioral features of Twitter, we are able to increase the aver- age, unsupervised F1 score by 21.52 points over a lexical baseline alone. 1 Introduction The importance of understanding political dis- course on social media platforms is becoming in- creasingly clear. In recent U.S. presidential elec- tions, Twitter was widely used by all candidates to promote their agenda, interact with supporters, and attack their opponents. Social interactions on such platforms allow politicians to quickly react to current events and gauge interest in and sup- port for their actions. These dynamic settings em- phasize the importance of constructing automated tools for analyzing this content. However, these same dynamics make constructing such tools dif- ficult, as the language used to discuss new events and political agendas continuously changes. Con- sequently, the rich social interactions on Twitter can be leveraged to help support such analysis by providing alternatives to direct supervision. In this paper we focus on political framing, a very nuanced political discourse analysis task, on a variety of issues frequently discussed on Twit- ter. Framing (Entman, 1993; Chong and Druck- man, 2007) is employed by politicians to bias the discussion towards their stance by emphasiz- ing specific aspects of the issue. For example, the debate around increasing the minimum wage can be framed as a quality of life issue or as an economic issue. While the first frame supports increasing minimum wage because it improves workers’ lives, the second frame, by conversely emphasizing the costs involved, opposes the in- crease. Using framing to analyze political dis- course has gathered significant interest over the last few years (Tsur et al., 2015; Card et al., 2015; Baumer et al., 2015) as a way to automatically an- alyze political discourse in congressional speeches and political news articles. Different from previ- ous works which focus on these longer texts or sin- gle issues, our dataset includes tweets authored by all members of the U.S. Congress from both par- ties, dealing with several policy issues …

Content truncated to 500 words