I have a big text corpus (documentation from a company) and I want to extract the terms that are specific to that area/business. I can do that using TF or TF-IDF and guide myself by the frequency of the words, which isn't always reliable.
I want to also do that for single, shorter sentences, but I think this is already more difficult. I was also thinking of using Wikipedia articles to train a model and then apply it to my documentation texts.
Is there any way of identifying words that are related to a specific field?