TF-IDF scores words by frequency in a document and rarity across documents.

It highlights terms that are important to a document but not common everywhere, making it useful for search and classic text classification.