Abstract
Le and Fokkens (2015) recently showed that taxonomy-based approaches are more reliable than corpus-based approaches in estimating human similarity ratings. On the other hand, distributional models provide much better coverage. The lack of an established similarity metric for adjectives in WordNet is a case in point. I present initial work to establish such a metric, and propose ways to move forward by looking at extensions to WordNet. I show that the shortest path distance between derivationally related forms provides a reliable estimate of adjective similarity. Furthermore, I find that a hybrid method combining this measure with vector-based similarity estimations gives us the best of both worlds: more reliable similarity estimations than vectors alone, but with the same coverage as corpus-based methods.- Anthology ID:
- 2016.gwc-1.58
- Volume:
- Proceedings of the 8th Global WordNet Conference (GWC)
- Month:
- 27--30 January
- Year:
- 2016
- Address:
- Bucharest, Romania
- Editors:
- Christiane Fellbaum, Piek Vossen, Verginica Barbu Mititelu, Corina Forascu
- Venue:
- GWC
- SIG:
- SIGLEX
- Publisher:
- Global Wordnet Association
- Note:
- Pages:
- 419–423
- Language:
- URL:
- https://aclanthology.org/2016.gwc-1.58
- DOI:
- Cite (ACL):
- Emiel van Miltenburg. 2016. WordNet-based similarity metrics for adjectives. In Proceedings of the 8th Global WordNet Conference (GWC), pages 419–423, Bucharest, Romania. Global Wordnet Association.
- Cite (Informal):
- WordNet-based similarity metrics for adjectives (van Miltenburg, GWC 2016)
- PDF:
- https://preview.aclanthology.org/improve-issue-templates/2016.gwc-1.58.pdf