Anonymous
TF-IDF is an approach based off ngrams. TF or term frequency is the count or ratio of words in a text blob. IDF or inverse document frequency finds the importance of the word across multiple text blobs. Together, they will be used to find important words in a text and their intersity depending on how many times it has been used. TF-IDF is used in multiple applications like text classification etc. The downside of the approach is that the vector that it creates is very sparse and as it's an ngram based approach it does not utilize context information.