Standardization

noun

say it: “stan-der-dih-ZAY-shun” /ˌstændərdəˈzeɪʃən/

Rescaling numbers so their average becomes 0 and their spread, the standard deviation, becomes 1.

What is standardization of a feature?

Standardization rescales a feature so its mean is 0 and its standard deviation is 1: subtract the mean, then divide by the standard deviation. A value then says how many standard deviations it is from average, whatever its units. scikit-learn's StandardScaler does it.

import numpy as np
spend = np.array([10, 20, 30])   # dollars per visit
print(((spend - spend.mean()) / spend.std()).round(2))   # [-1.22  0.    1.22]

In a sentence

  • “After standardization, a spend of $30 became 1.22: a little over one standard deviation above average.”
  • “Standardization let age and income count equally in the distance.”

Level: Journeyman · More AI Basics terms at this level

Updated