Clever Geek Handbook

Natural language processing - like ngram

I am currently working on an NLP project that tries to distinguish synonyms (derived from Python NLTK using WordNet) in context. I've explored a lot of NLP concepts trying to find what I want and the closest I've found is n-grams, but that's not really perfect.

Suppose I am trying to find the correct definition for the verb "box". "Box" can mean "fight" or "package"; however, the word "ring" or "fighter" appears elsewhere in the text. As far as I understand, n-gram will be "box box" or "box ring", which is pretty funny as a phrase and is unlikely to appear. But on a concept map, the action of "box" can be associated with a "ring" because they are conceptually related.

I-gram, what do I want? Is there another name for this? Any help finding the relational data you're looking for?

All help is appreciated.

+3


source to share


1 answer


You might want to explore word sense disambiguation (WSD)

this is the problem of determining what "sense" (meaning) of a word is activated by the use of a word in a particular context, a process that appears to be largely unconscious in humans.



+2


source






More articles:


All Articles