TfidfVectorizer

A Scikit-learn utility that converts raw text into a matrix of TF-IDF features for machine learning models.

TfidfVectorizer combines CountVectorizer and TfidfTransformer into a single, high-performance class. It tokenizes documents, calculates term frequency (TF), and applies inverse document frequency (IDF) scaling to down-weight common stop words (like 'the' or 'is') while highlighting unique, informative terms. This process transforms unstructured text into a sparse NumPy matrix compatible with estimators like LogisticRegression or LinearSVC. Key parameters include 'max_features' to limit vocabulary size and 'ngram_range' to capture multi-word phrases (e.g., 'machine learning').

https://scikit-learn.org/stable/modules/generated/sklearn.feature_extraction.text.TfidfVectorizer.html