Difference between revisions of "Controversial events detection"

From Cohen Courses
Jump to navigationJump to search
m
m
Line 1: Line 1:
 
= Controversial event detection =
 
= Controversial event detection =
  
−
== Team members =
+
== Team members ==
  
 
== Project idea ==
 
== Project idea ==
Line 9: Line 9:
 
Controversy-wise, Christmas day is relatively one sided, with most of the text mentioning it being relatively homogeneous.
 
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).
 
In contrast, the Presidential debates event will have obvious sides (supporting the different candidates).
−
Hence, we define
+
We believe that controversial events will have higher "entropy" in the textual content describing it.
−
Much prior work has focused on event detection (todo: cite).
 
−
There has also been some work on identifying controversial events (todo: cite).
 
  
 +
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.
  
−
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
 
  
 
== Data ==
 
== Data ==

Revision as of 21:06, 1 October 2012

Controversial event 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). We believe that controversial events will have higher "entropy" in the textual content describing it.

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.


Data

Related work

Related materials

Retrieved from "http://curtis.ml.cmu.edu/w/courses/index.php?title=Controversial_events_detection&oldid=12905"