SimpleImputer
noun
say it: “SIM-pul im-PYOO-ter” /ˈsɪmpəl ɪmˈpjutər/
A scikit-learn tool that fills in missing values in each column with one value learned from the training data, such as the median.
What is SimpleImputer?
SimpleImputer is scikit-learn's tool for filling missing values: fit() learns one value per column from the
training rows, such as the median (strategy="median") or the most common value, and transform() fills the gaps
with it. Inside a pipeline, the test rows never shape what it learns.
import numpy as np
from sklearn.impute import SimpleImputer
ages = [[2], [4], [np.nan], [10]] # years; one is missing
print(SimpleImputer(strategy="median").fit_transform(ages).ravel()) # [ 2. 4. 4. 10.]
In a sentence
- “Putting the SimpleImputer in the pipeline meant the test rows never shaped its median.”
- “The SimpleImputer filled each missing age with the training median of 4 years.”
Level: Journeyman · More Python terms at this level
Updated