Controversial events detection

= Team members

Project idea

In our project, we propose to jointly detect events and the controversy surrounding it in the context of social media. For example, Christmas day is an event that receives the most attention around December 25th, while the Presidential debates once every four years. Controversy-wise, Christmas day is relatively one sided, with most of the text mentioning it being relatively homogeneous. In contrast, the Presidential debates event will have obvious sides (supporting the different candidates). Hence, we define Much prior work has focused on event detection (todo: cite). There has also been some work on identifying controversial events (todo: cite).

We propose in this project to look at the automatic detection of events in social media that are explicitly controversial in nature. While much prior work has investigated event detection in social media text streams in a non-probabilistic context (e.g., detecting “bursty” words and clustering them together) [5], [2], we focus on events that are interesting by virtue of their controversy. For example, while New Year’s Day is an event that spikes each year around January 1, it is relatively uninteresting since the text describing it is more or less homogeneous (e.g., “Happy New Year!”); in contrast, the event described by the Susan G. Komen Foundation decision to pull support for Planned Parenthood spiked on February 3, 2012; this event has clear sides (one supporting the decision and one opposing it) who discuss it in radically different ways. Recent work has begun looking at detecting controversy in social media [4], but has been limited to simple event classification; our goal is not only to detect controversial events but also to discover what the different sides are – both grouping the individuals associated with each faction and describing how each faction talks about the event differently. With this goal, graphical models are a natural solution