Comparison of the Baseline Knowledge-, Corpus-, and Web-based Similarity Measures for Semantic Relations Extraction

Panchenko, Alexander;Fairon, Cédrick
(2011) GEMS 2011 Workshop on Geometrical Models of Natural Language Semantics, EMNLP 2011, — Location: Edinbourgh, Angleterre

Files

W11-2502.pdf
  • Open Access
  • Adobe PDF
  • 1.07 MB
Download

Details

Authors
Abstract
Unsupervised methods of semantic relations extraction rely on a similarity measure be- tween lexical units. Similarity measures differ both in kinds of information they use and in the ways how this information is transformed into a similarity score. This paper is making a step further in the evaluation of the avail- able similarity measures within the context of semantic relation extraction. We compare 21 baseline measures – 8 knowledge-based, 4 corpus-based, and 9 web-based metrics with the BLESS dataset. Our results show that existing similarity measures provide signifi- cantly different results, both in general per- formances and in relation distributions. We conclude that the results suggest developing a combined similarity measure.
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Panchenko, A. (2011). Comparison of the Baseline Knowledge-, Corpus-, and Web-based Similarity Measures for Semantic Relations Extraction. GEMS 2011 Workshop on Geometrical Models of Natural Language Semantics, EMNLP 2011, Edinbourgh, Angleterre. https://hdl.handle.net/2078.5/159963