Stopwords are already getting filtered in the training data. So, even if we use n-grams in CountVectorizer, we cannot train the model on stopwords such as ‘not’ etc.
Using n-grams will not help in getting right prediction
Hello Shubham,
The concept of stopwords is completely different as that from n-grams. The purpose of using n-grams is to collectively use ‘n’ number of words in our vocabulary because they make very little sense when used individually and only represent the right context when used together.
Our goal of removing stopwords was to eliminate the words which do not represent much of a context for us.