I have a corpus and am looking to find whether there is an antonymy relationship within each sentence, using Wordnet.
Does anyone have a clue whether looking for antonyms in Wordnet is at all possible? Thank you!
Antonyms apply at a lemma level, not at a synset level. (According to documentation and this issue in the nltk google code project.) Using an example from the latter link:
>>> wordnet.synset('light.a.01').antonyms()
[]
>>> wordnet.lemma('light.a.01.light').antonyms()
[Lemma('heavy.a.01.heavy')]