Implementation:Scikit learn Scikit learn IterativeImputer

Knowledge Sources	Scikit_learn Scikit-learn Docs
Domains	Machine Learning, Missing Data, Imputation
Last Updated	2026-02-08 15:00 GMT

Overview

Concrete implementation of multivariate iterative imputation for missing values provided by scikit-learn.

Description

The IterativeImputer class implements a strategy for imputing missing values by modeling each feature with missing values as a function of other features in a round-robin fashion. At each step, a feature is designated as output and the other features are treated as inputs. An estimator is fit on known values and used to predict missing ones. This process repeats for multiple iterations. The class is experimental and requires explicit import from sklearn.experimental.

Usage

Use IterativeImputer when you need sophisticated imputation that accounts for feature correlations, rather than simple mean/median imputation. It is particularly useful when missing data patterns are not completely at random (MCAR).

Code Reference

Source Location

• Repository: scikit-learn
• File: sklearn/impute/_iterative.py

Signature

class IterativeImputer(_BaseImputer):
    def __init__(
        self,
        estimator=None,
        *,
        missing_values=np.nan,
        sample_posterior=False,
        max_iter=10,
        tol=1e-3,
        n_nearest_features=None,
        initial_strategy="mean",
        fill_value=None,
        imputation_order="ascending",
        skip_complete=False,
        min_value=-np.inf,
        max_value=np.inf,
        verbose=0,
        random_state=None,
        add_indicator=False,
        keep_empty_features=False,
    ):
        ...

    def fit(self, X, y=None, **fit_params):
        ...

    def transform(self, X):
        ...

    def fit_transform(self, X, y=None, **fit_params):
        ...

Import

from sklearn.experimental import enable_iterative_imputer  # noqa
from sklearn.impute import IterativeImputer

I/O Contract

Inputs

Name	Type	Required	Description
X	array-like of shape (n_samples, n_features)	Yes	Data with missing values to impute
estimator	estimator instance	No	Estimator to use for imputation (default: BayesianRidge)
max_iter	int	No	Maximum number of imputation rounds (default: 10)
initial_strategy	str	No	Strategy for initial imputation: "mean", "median", "most_frequent", "constant"
imputation_order	str	No	Order of feature imputation: "ascending", "descending", "roman", "arabic", "random"
missing_values	int, float, or np.nan	No	Placeholder for missing values

Outputs

Name	Type	Description
X_imputed	ndarray of shape (n_samples, n_features)	Data with missing values imputed
n_iter_	int	Number of rounds of imputation performed

Usage Examples

Basic Usage

import numpy as np
from sklearn.experimental import enable_iterative_imputer  # noqa
from sklearn.impute import IterativeImputer

X = np.array([[1, 2], [np.nan, 3], [7, 6], [4, np.nan]])
imp = IterativeImputer(max_iter=10, random_state=42)
X_imputed = imp.fit_transform(X)
print(X_imputed)

Related Pages

• Principle:Scikit_learn_Scikit_learn_Missing_Data_Imputation