Comparison of the results when training on one type of dataset or domain and testing on another.

<p>Coreference resolution results comparison on ACE2004 (i.e., A), CoNLL2012 (i.e., C) and SemEval2010 newswire (i.e., NW) and broadcast news (i.e., BN) datasets. Each column represents a model trained on a specific dataset, while each row represents a dataset. Values represent -scores of MUC <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0100101#pone.0100101-Vilain1" target="_blank">[9]</a>, BCubed <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0100101#pone.0100101-Bagga1" target="_blank">[41]</a> and CEAF <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0100101#pone.0100101-Luo3" target="_blank">[42]</a>, respectively.