This tarball contains the data accompanying the paper:

Robust Systems for Preposition Error Correction Using Wikipedia Revisions
Aoife Cahill; Nitin Madnani; Joel Tetreault; Diane Napolitano
Proceedings of NAACL, 2013.

The tarball contains two files:

1. 'wiki.prep.sents' contains the part of speech tagged sentences from Wikipedia containing the preposition errors. The part-of-speech
    tags were generated using the Stanford Tagger using the "wsj-0-18-left3words-distsim.tagger" model.
2. 'wiki.prep.corrections' contains the stand-off corrections corresponding to each sentence. It contains the following fields in order:

    - The token offset of the preposition error in the sentence (numbered starting from 1)
    - The incorrect preposition
    - The part-of-speech tag of the incorrect preposition (IN or TO)
    - The correct preposition
    - The part-of-speech tag of the correct preposition (IN or TO)
    - Wikipedia Article ID containing the sentence
    - The ID of the article revision containing the preposition error
    - The ID of the article revision containing the preposition correction
    - A binary field indicating whether the correction was derived by collapsing an edit chain or not
    - A field indicating whether the correction is clean, somewhat clean or dirty (defined in the paper)
