A B-Spline Hidden Markov Clustering Model with an Application to Wikipedia Elite Editor Activities Modeling

• Activity Number: 552
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Wutao Wei*
• Keywords: Hidden Markov Model ; B-spline ; Wikipedia ; clustering

Abstract

Wikipedia has over 20 million editors now. However, the work mode of different editors are entirely different. Over 90% of the materials are contributed by less than 1% of the editors. As a result, we proposed a B-spline Hidden Markov model to model the longitudinal activities for each editor, and clustered them into different groups according to their work mode similarity. We assume each editor has different work states, which indicate different work patterns. In each pattern, a set of B-spline functions will be used for representing the user activities. In the end, we are able to tell the pattern of elite editors and their road to become an elite.