Oct-06-2021, 07:57 AM
(This post was last modified: Oct-06-2021, 07:57 AM by AlekseyPython.)
With word2vec I can vectorize both individual words and whole texts (as the arithmetic mean of all words included in the text). But at the same time, positive and negative text receive one vector representation, although they talk about mutually opposite things. Searching for the particles "no" or "not" in the text to reveal negation is not always effective, because many words have antonyms and such pairs are not recognized as opposites.
How to recognize negation in a text?
How to recognize negation in a text?