Implementation:Scikit learn Scikit learn SparseFuncs

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

Overview

Concrete utility module for efficient operations on sparse matrices provided by scikit-learn.

Description

The sparsefuncs module provides a collection of utilities for working with sparse CSR and CSC matrices. It includes inplace scaling operations (inplace_csr_column_scale, inplace_csr_row_scale), mean/variance computation along axes, min/max operations, sparse matrix multiplication to dense output, and count-of-nonzero functions. These operations avoid materializing full dense matrices.

Usage

Use these functions when you need to perform feature-wise or sample-wise operations on sparse matrices without converting them to dense format, such as during feature scaling, normalization, or statistical computation.

Code Reference

Source Location

• Repository: scikit-learn
• File: sklearn/utils/sparsefuncs.py

Signature

def inplace_csr_column_scale(X, scale):
    ...

def inplace_csr_row_scale(X, scale):
    ...

def mean_variance_axis(X, axis, weights=None, return_sum_weights=False):
    ...

def incr_mean_variance_axis(X, *, axis, last_mean, last_var, last_n, weights=None):
    ...

def min_max_axis(X, axis, ignore_nan=False):
    ...

def count_nonzero(X, axis=None, sample_weight=None):
    ...

def sparse_matmul_to_dense(A, B):
    ...

Import

from sklearn.utils.sparsefuncs import (
    inplace_csr_column_scale,
    mean_variance_axis,
    count_nonzero,
)

I/O Contract

Inputs

Name	Type	Required	Description
X	sparse matrix (CSR or CSC)	Yes	Sparse matrix to operate on
scale	ndarray	Yes	Scale factors for inplace scaling operations
axis	int	Yes	Axis along which to compute (0 for columns, 1 for rows)
weights	ndarray	No	Sample weights for weighted computations

Outputs

Name	Type	Description
means	ndarray	Mean values along the specified axis
variances	ndarray	Variance values along the specified axis
mins	ndarray	Minimum values along the specified axis
maxs	ndarray	Maximum values along the specified axis

Usage Examples

Basic Usage

import numpy as np
from scipy import sparse
from sklearn.utils.sparsefuncs import inplace_csr_column_scale, mean_variance_axis

# Create a sparse matrix
X = sparse.random(100, 10, density=0.3, format="csr")

# Compute mean and variance along columns
means, variances = mean_variance_axis(X, axis=0)
print(means.shape)  # (10,)

# Inplace column scaling
scale = np.array([2.0] * 10)
inplace_csr_column_scale(X, scale)

Related Pages

• Principle:Scikit_learn_Scikit_learn_Dimensionality_Reduction