What is the difference between Transformer and Estimator in Scikit-Learn?

Estimators learn from data using fit(), while transformers additionally modify data using transform().

In Scikit-Learn’s API design, an estimator is any object that learns from data via fit(). A transformer is a specialized estimator that implements transform() (and often fit_transform()). Transformers are used in preprocessing pipelines, while estimators include predictive models like classifiers and regressors. This distinction is central to Scikit-Learn’s composability philosophy.