Is it always necessary to either stem/lemmatize words when working with TF-IDF?

306 Views Asked by At

I'm using TF-IDF along with cosine similarity in order to compute document similarity. I was wondering if it's always necessary to stem/lemmatize the words in the document. Are there times where based on the task, it's better not to stem/lemmatize?

0

There are 0 best solutions below