Graph-Based Breaking News Detection on Wikipedia

Event detection in social media usually exploits information from social-networking platforms, such as Twitter or Facebook. However, previous research has suggested that Wikipedia constitutes a valuable source of information for the task of detecting breaking news. In this work we adapt a graph-based algorithm to the Wikipedia context, and compare it to the state-of-The-Art Wikipedia real-Time monitoring method. The main idea behind the proposed method is to extract breaking news by looking at unusual activity in the Wikipedia edit stream. We assess the performance of the two competing algorithms by measuring the percentage of true events correctly identified. Results show that the proposed graph-based method achieves better accuracy and coverage. Wikipedia has been largely recognised as a valuable source for detecting breaking news (Osborne et al. 2012). However, to the best of our knowledge, existing works are just based on spike-detection approaches that look at the number of page views or revisions of an article. In this work we propose a novel method consisting of an adaptation to the Wikipedia context of a graph-based approach, which has been traditionally used for detecting events from online user- generated content