PAN 2012

From WikiPapers
Jump to: navigation, search

PAN 2012: Competition on Quality Flaw Prediction in Wikipedia is a contest. The competition is part of the PAN 2012 Lab, held in conjunction with the CLEF'12 conference in Rome, Italy.

The previous PAN Labs have addressed quality issues in Wikipedia in the form of vandalism. However, the majority of quality flaws is not caused due to malicious intentions but stem from edits by inexperienced authors; examples include poor writing style, unreferenced statements, or missing neutrality. This year, the organizers generalize the vandalism detection task and focus on the prediction of particular quality flaws in Wikipedia articles.


Title Author(s) Keyword(s) Language Abstract R C
FlawFinder: A Modular System for Predicting Quality Flaws in Wikipedia Oliver Ferschke
Iryna Gurevych
Marc Rittberger
English With over 23 million articles in 285 languages, Wikipedia is the largest free knowledge base on the web. Due to its open nature, everybody is allowed to access and edit the contents of this huge encyclopedia. As a downside of this open access policy, quality assessment of the content becomes a critical issue and is hardly manageable without computational assistance. In this paper, we present FlawFinder, a modular system for automatically predicting quality flaws in unseen Wikipedia articles. It competed in the inaugural edition of the Quality Flaw Prediction Task at the PAN Challenge 2012 and achieved the best precision of all systems and the second place in terms of recall and F1-score. 0 1
On the Use of PU Learning for Quality Flaw Prediction in Wikipedia Edgardo Ferretti
Donato Hernández Fusilier
Rafael Guzmán Cabrera
Manuel Montes y Gómez
Marcelo Errecalde
Paolo Rosso
English In this article we describe a new approach to assess Quality Flaw Prediction in Wikipedia. The partially supervised method studied, called PU Learning, has been successfully applied in classifications tasks with traditional corpora like Reuters-21578 or 20-Newsgroups. To the best of our knowledge, this is the first time that it is applied in this domain. Throughout this paper, we describe how the original PU Learning approach was evaluated for assessing quality flaws and the modifications introduced to get a quality flaws predictor which obtained the best F1 scores in the task “Quality Flaw Prediction in Wikipedia” of the PAN challenge. 0 1