Closed JherezTaylor closed 7 years ago
ConceptNet can be used to yield the co-occurrence of words for each word in the HS bag See https://github.com/commonsense/conceptnet5
This was fixed by d399708 but it's slow and should be used sparingly, the better approach is the use the brown cluster id
as a measure of similarity.
k
similar words