More about HKUST
The Politics of Comments: Predicting Political Orientation of News Stories with Commentators' Sentiment Patterns
Speaker: Souneil Park KAIST, Korea Title: "The Politics of Comments: Predicting Political Orientation of News Stories with Commentators' Sentiment Patterns" Date: Friday, 25 March 2011 Time: 10:30am - 11:30am Venue: Room 2612A (via lifts 31/32), HKUST Abstract: Political views frequently conflict in the coverage of contentious political issues, potentially causing serious social problems. While much useful to users, it is very difficult to analyse news articles and identify the political views of news articles. We develop an important method which identifies political orientation of news stories without complex news text analysis. Admitting the practical limitation of news text analysis, the method takes advantage of commentators' participation as well as their knowledge and intelligence; their interpretation of the political orientation is condensed in the sentiment of comments. Based on extensive study on commentators' behaviors, it further notes and verifies that there exist predictive commentators with a clear political preference and that they are highly likely to show their views consistently towards various political issues. They actively express their preferences responding to the political news articles, showing a high degree of regularity in their sentiment patterns. The sentiment expressed in their comments is a strong indicator of the political orientation of an article. Also, comments are usually concise, often express sentiments explicitly, and easier to analyze. The use of these predictive commentators and their comments greatly reduces the high complexity of political view identification. This work is published in ACM CSCW 2011. ***************** Biography: Souneil Park is a Ph.D candidate in KAIST, Daejeon, Korea. He currently conducts interdisciplinary research combining mass communication with computer science. His research interest includes Internet journalism, social media applications, web science, and human computer interaction.