Evaluating Event Credibility on Twitter Presented by YananXie College of Computer Science, Zhejiang University 2012
Motivation Since its launch in 2006, Twitter has gained huge popularity. 4% of the massive number of tweets constitute news. Recent surveys show that just ~8% people trust Twitter, while just ~12% trust their friends’ Twitter streams. A lot of rumors have been spread using Twitter in the recent past and have resulted into a lot of damage
Problem Definition Example: A bull jumped into the crowds in a Spanish arena injuring 30 people.
Classification-Based Approach Lacks: With high probability, credible users provide credible tweets; Average credibility of tweets is higher for credible events than that for non- credible events; and Events sharing many common words and topics should have similar credibility scores.
Performing Event Graph Optimization similar events get similar credibility scores, and change in event credibility vector is controlled.