Wikipedia Vandalism Detection Through Machine Learning: Feature Review and New Proposals: Lab Report for PAN at CLEF 2010

TL;DR

This paper reviews features for detecting Wikipedia vandalism using machine learning, extends previous frameworks, and reports that a Random Forest classifier achieved top performance in the PAN 2010 vandalism detection challenge.

Contribution

It extends prior vandalism detection frameworks by proposing new features and demonstrates that Random Forest classifiers outperform others in this task.

Findings

Random Forest achieved an AUC of 0.92236

The approach ranked first in the PAN 2010 vandalism detection task

Supervised learning effectively detects vandalism in Wikipedia edits

Abstract

Wikipedia is an online encyclopedia that anyone can edit. In this open model, some people edits with the intent of harming the integrity of Wikipedia. This is known as vandalism. We extend the framework presented in (Potthast, Stein, and Gerling, 2008) for Wikipedia vandalism detection. In this approach, several vandalism indicating features are extracted from edits in a vandalism corpus and are fed to a supervised learning algorithm. The best performing classifiers were LogitBoost and Random Forest. Our classifier, a Random Forest, obtained an AUC of 0.92236, ranking in the first place of the PAN'10 Wikipedia vandalism detection task.