Hello, I want to weight terms of a large text based on their semantics (not on their frequency (TF-IDF)). Is there a way to do that using NLTK or other means? through a vectorizer? For example: a certain term weights more than others etc... Thanks