fit()

noun

say it: “FIT” /ˈfɪt/

In scikit-learn, the method that trains a model: it learns from examples and their right answers.

What does fit() do in scikit-learn?

fit() is the scikit-learn method that trains a model: model.fit(X, y) learns from the features X (one row per example) and the labels y (the right answers). It returns the model, so .predict() can follow on the same line.

from sklearn.linear_model import LinearRegression
temps = [[60], [70], [80]]   # features: one row per day
cups = [20, 40, 60]          # labels: cups of lemonade sold
model = LinearRegression().fit(temps, cups)
print(model.coef_)   # [2.]: 2 more cups per degree

In a sentence

  • “A model can't predict until you've called fit().”
  • “Calling fit() on last summer's sales trained the lemonade model.”

Level: Apprentice · More Python terms at this level

Updated