I am a student who is studying machine running alone. As far as I know, the Tf-idf Vectorizer is a function that combines the Count Vectorizer and the Tf-idf Transformer. I was wondering if the Tf-idf Transformer process could actually have a impact on the model learning process. In my experience, there seemed to be no model performance difference between using CountVectorizer and Tf-idf Vectorizer in preprocessing String vector. Is there a situation where we need to consider the Tf-idfVectorizer rather than the CountVectorizer when processing a String Vector?

  • $\begingroup$ If you're asking about a particular library, that's off-topic here. "Tf-idf Vectorizer" might refer to some specific code in some particular machine learning library, but without context it isn't clear what you mean by that. $\endgroup$ – D.W. Feb 10 at 6:22
  • $\begingroup$ Anyway, yes, in some contexts TF-IDF does offer improvements. See e.g. en.wikipedia.org/wiki/Tf%E2%80%93idf. $\endgroup$ – D.W. Feb 10 at 6:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.