Kiran Garimella and colleagues at Aalto University in Finland say they have developed a more reliable way of spotting controversies in the Twitterstream in real time.
Although other researchers have focused on pre-identified arguments, Garimella's team sought to determine whether the structure of controversial conversations is different from that of benign discussions. The researchers theorized this structure could be spotted by studying the network of connections between those involved in a topic, the structure of endorsement (or who agrees with whom), and the sentiment of the discussion.
They tested their theory by first studying 10 conversations associated with hashtags known to be controversial and 10 known to be benign. The researchers then mapped out the structure of the discussion by examining the networks of retweets, follows, and keywords.
The team says the retweet and follow graphs reveal clusters that might indicate a form of polarization, as networks can be partitioned into groups on either side of the issue. Images of controversies resemble fireworks, and the team says the method can spot controversies better than existing tools.
From Technology Review
View Full Article
Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA
No entries found