Researchers at the University of Alabama at Birmingham have developed a tool that could be used to uncover phony Wikipedia contributors.
Companies often pay people to set up Wikipedia accounts and edit articles in their favor, and the algorithm is designed to analyze the way articles are written and spot if they are edited by the same person. Ragib Hasan and colleagues trained an algorithm on the editing notes for more than 600 investigations by Wikipedia into fake accounts, also known as sock puppet accounts. The algorithm scanned about 230 features of the writing, such as grammatical quirks, and could predict which accounts were puppet accounts with a 75-percent accuracy rate, defined as agreeing with the conclusion of Wikipedia investigators.
"Sock-puppet investigations are incredibly time-consuming for Wikipedia editors, so anything that can help reduce the workload should be welcome," Hasan says.
From New Scientist
View Full Article - May Require Registration
Abstracts Copyright © 2013 Information Inc., Bethesda, Maryland, USA
No entries found