The PAN Wikipedia vandalism corpus 2011 (PAN-WVC-11) is a corpus for the evaluation of automatic vandalism detectors for Wikipedia. For research purposes the corpus can be used free of charge.

This corpus supplements the PAN-WVC-10, which features only English edits. Both corpora should be used to get more representative results.


To download the corpus use the following link:

If you use the dataset in your research, please send us a copy of your publication. We kindly ask you to refer to the corpus via [bib].


As part of our research on automatic vandalism detection we have compiled a corpus of vandalism cases found in Wikipedia. The corpus compiles 29949 edits on 24351 Wikipedia articles, among which 2813 vandalism edits have been identified. The corpus features 9985 English edits, 9990 German edits, and 9974 Spanish edits. To annotate the corpus we have used Amazon's Mechanical Turk; each edit was presented to a number of annotators who were asked to decide whether it is vandalism or regular, and the agreement of the annotators was analyzed in order to label an edit.

The corpus has been successfully employed in the 2nd International Competition on Wikipedia Vandalism Detection, PAN'11, which was held in conjunction with the CLEF'11 conference.

Previous Corpus Versions. This corpus supplements the PAN-WVC-10 corpus which consists of more than 30000 English article edits. While the edits for both corpora have been chosen from the same time frame, they do not intersect, so that both corpora may be combined to allow for a more representative evaluation.


Students: Teresa Holfeld