Estimator

noun

say it: “ES-tih-may-ter” /ˈɛstəˌmeɪtər/

In scikit-learn, any object that learns from data through a fit method, such as a model or a tool that prepares data.

What is an estimator in scikit-learn?

An estimator is scikit-learn's word for any object that learns from data with a fit() method: a model such as LinearRegression, and also a tool such as StandardScaler. Every estimator works the same way, so swapping one for another changes one line.

from sklearn.linear_model import LinearRegression
from sklearn.tree import DecisionTreeRegressor
for model in (LinearRegression(), DecisionTreeRegressor()):
    model.fit([[60], [70], [80]], [20, 40, 60])   # the same call for both
    print(model.predict([[70]]))   # [40.] each time

In a sentence

  • “She swapped one estimator for another and the rest of the code stayed the same.”
  • “Every estimator in scikit-learn learns with fit().”

Level: Journeyman · More Python terms at this level

Updated