Quelle est la meilleure façon de comprendre la similitude sémantique des mots? Word2Vec est correct, mais pas idéal:
# Using the 840B word Common Crawl GloVe vectors with gensim:
# 'hot' is closer to 'cold' than 'warm'
In [7]: model.similarity('hot', 'cold')
Out[7]: 0.59720456121072973
In [8]: model.similarity('hot', 'warm')
Out[8]: 0.56784095376659627
# Cold is much closer to 'hot' than 'popular'
In [9]: model.similarity('hot', 'popular')
Out[9]: 0.33708479049537632
Les méthodes Wordnet de NLTK semblent simplement abandonner:
In [25]: print wn.synset('hot.a.01').path_similarity(wn.synset('warm.a.01'))
None
Quelles sont les autres options?