Implementation:Scikit learn Scikit learn IncrementalPCA

Knowledge Sources	Scikit_learn Scikit-learn Docs
Domains	Dimensionality Reduction, Online Learning
Last Updated	2026-02-08 15:00 GMT

Overview

Concrete tool for memory-efficient incremental principal components analysis provided by scikit-learn.

Description

IncrementalPCA performs linear dimensionality reduction using Singular Value Decomposition of the data in mini-batches, keeping only the most significant singular vectors. Unlike standard PCA, it has constant memory complexity on the order of batch_size times n_features, enabling use with np.memmap files and sparse input without loading the entire dataset into memory. The data is centered but not scaled before applying the SVD.

Usage

Use IncrementalPCA when your dataset is too large to fit in memory for standard PCA, or when data arrives in a streaming fashion. It is particularly well suited for out-of-core PCA on large datasets, processing data from memmap files, and scenarios requiring partial_fit for online learning.

Code Reference

Source Location

Repository: scikit-learn
File: sklearn/decomposition/_incremental_pca.py

Signature

class IncrementalPCA(_BasePCA):
    def __init__(
        self,
        n_components=None,
        *,
        whiten=False,
        copy=True,
        batch_size=None,
    ):

Import

from sklearn.decomposition import IncrementalPCA

I/O Contract

Inputs

Name	Type	Required	Description
n_components	int	No	Number of components to keep. Defaults to min(n_samples, n_features) if None.
whiten	bool	No	When True, component vectors are scaled to ensure uncorrelated outputs with unit variance (default=False).
copy	bool	No	If False, X will be overwritten (default=True).
batch_size	int	No	Number of samples per batch. Defaults to 5 * n_features if None.

Outputs

Name	Type	Description
components_	ndarray of shape (n_components, n_features)	Principal axes in feature space (directions of maximum variance).
explained_variance_	ndarray of shape (n_components,)	Variance explained by each selected component.
explained_variance_ratio_	ndarray of shape (n_components,)	Proportion of variance explained by each selected component.
singular_values_	ndarray of shape (n_components,)	Singular values corresponding to each selected component.
mean_	ndarray of shape (n_features,)	Per-feature empirical mean, estimated from the training set.
n_samples_seen_	int	Total number of samples processed by the estimator.

Usage Examples

Basic Usage

import numpy as np
from sklearn.decomposition import IncrementalPCA

X = np.random.rand(1000, 50)
ipca = IncrementalPCA(n_components=10, batch_size=200)
X_transformed = ipca.fit_transform(X)
print(X_transformed.shape)  # (1000, 10)

# Alternatively, use partial_fit for streaming data
ipca2 = IncrementalPCA(n_components=10)
for batch in np.array_split(X, 5):
    ipca2.partial_fit(batch)
X_transformed2 = ipca2.transform(X)
print(X_transformed2.shape)  # (1000, 10)

Related Pages

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment