Closed nicolaierbs closed 7 years ago
citing 2014 comments from Torsten here (translated):
This metric would fit into the module lexical-asl, where Jaccard and Containment are already available. Dice should require just a small change (is 2*Jaccard). Maybe do some refactoring, i.e. base class instead of override.
the weighted combination of ngrams across different levels (lemma, POS, token) is also applicable in a generic way. Therefore it might be better to create a wrapper-similarity class which then can be used for other ngram-based similarity measures as well. Something like: new WeightedCombinationMeasure(new WordNgramDiceMeasure(), 2, 3, 4)
If there is a need for this measure, please re-open.
Original issue 34 created by dkpro on 2014-11-21T08:44:14.000Z:
Add the sentence/pattern similarity metric used in the paper
Automated Verb Sense Labelling Based on Linked Lexical Resources Kostadin Cholakov and Judith Eckle-Kohler and Iryna Gurevych In: Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics (EACL 2014), p. 68-77, Association for Computational Linguistics, April 2014.
to lexical-asl