Implementation:Scikit learn Scikit learn L1MinC

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

Overview

Concrete tool for computing the lowest bound for the regularization parameter C in L1-penalized classifiers, provided by scikit-learn.

Description

The l1_min_c function returns the lowest bound for the regularization parameter C such that for C in (l1_min_c, infinity) the model is guaranteed not to be empty. This applies to L1-penalized classifiers such as LinearSVC with penalty='l1' and LogisticRegression with l1_ratio=1. The computed value is valid when the class_weight parameter is not set.

Usage

Use this function before fitting an L1-penalized linear classifier to determine the minimum useful value of the C parameter, particularly when constructing regularization paths.

Code Reference

Source Location

• Repository: scikit-learn
• File: sklearn/svm/_bounds.py

Signature

@validate_params(...)
def l1_min_c(X, y, *, loss="squared_hinge", fit_intercept=True, intercept_scaling=1.0):

Import

from sklearn.svm import l1_min_c

I/O Contract

Inputs

Name	Type	Required	Description
X	array-like or sparse matrix of shape (n_samples, n_features)	Yes	Training feature matrix
y	array-like of shape (n_samples,)	Yes	Target vector
loss	str	No	Loss function: 'squared_hinge' or 'log' (default 'squared_hinge')
fit_intercept	bool	No	Whether the intercept is fitted (default True)
intercept_scaling	float	No	Scaling factor for synthetic intercept feature (default 1.0)

Outputs

Name	Type	Description
l1_min_c	float	The lowest bound for C guaranteeing a non-empty model

Usage Examples

Basic Usage

from sklearn.svm import l1_min_c
from sklearn.datasets import make_classification

X, y = make_classification(n_samples=100, n_features=20, random_state=0)
min_c = l1_min_c(X, y, loss="squared_hinge")
print(f"Minimum C value: {min_c:.6f}")

Related Pages

• Principle:Scikit_learn_Scikit_learn_Support_Vector_Machines