Implementation:Scikit learn Scikit learn BaseSVC

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 providing the base class infrastructure for all libsvm-based support vector machine estimators provided by scikit-learn.

Description

BaseSVC and BaseLibSVM are the abstract base classes for SVM classification and regression estimators in scikit-learn. BaseLibSVM wraps the libsvm library and provides the core fitting and prediction machinery including kernel computation, support vector extraction, and sparse data handling. BaseSVC extends it with classification-specific functionality including decision function shapes (one-vs-rest, one-vs-one), class weight handling, and probability calibration via Platt scaling. The module also provides utility functions for liblinear-based SVMs.

Usage

Use BaseSVC and BaseLibSVM as foundation classes when understanding the SVM implementation architecture. End users should use the concrete subclasses such as SVC, NuSVC, SVR, or LinearSVC instead of instantiating these base classes directly.

Code Reference

Source Location

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

Signature

class BaseLibSVM(BaseEstimator, metaclass=ABCMeta):
    def __init__(
        self,
        kernel,
        degree,
        gamma,
        coef0,
        tol,
        C,
        nu,
        epsilon,
        shrinking,
        probability,
        cache_size,
        class_weight,
        verbose,
        max_iter,
        random_state,
    ):

class BaseSVC(ClassifierMixin, BaseLibSVM, metaclass=ABCMeta):
    def __init__(
        self,
        kernel,
        degree,
        gamma,
        coef0,
        tol,
        C,
        nu,
        shrinking,
        probability,
        cache_size,
        class_weight,
        verbose,
        max_iter,
        decision_function_shape,
        random_state,
        break_ties,
    ):

Import

from sklearn.svm._base import BaseLibSVM, BaseSVC

I/O Contract

Inputs

Name	Type	Required	Description
kernel	str or callable	Yes	Kernel type: 'linear', 'poly', 'rbf', 'sigmoid', 'precomputed', or callable
degree	int	Yes	Degree of the polynomial kernel function
gamma	str or float	Yes	Kernel coefficient for 'rbf', 'poly', and 'sigmoid'
coef0	float	Yes	Independent term in kernel function
tol	float	Yes	Tolerance for stopping criterion
C	float	Yes	Regularization parameter
nu	float	Yes	Upper bound on training errors fraction (for Nu-SVM variants)
shrinking	bool	Yes	Whether to use the shrinking heuristic
probability	bool	Yes	Whether to enable probability estimates
cache_size	float	Yes	Size of the kernel cache in MB
class_weight	dict or str or None	Yes	Weights associated with classes
verbose	bool	Yes	Enable verbose output
max_iter	int	Yes	Maximum number of iterations (-1 for no limit)
decision_function_shape	str	Yes	Decision function shape: 'ovr' or 'ovo' (BaseSVC only)
random_state	int, RandomState, or None	Yes	Random state for probability estimation
break_ties	bool	Yes	Whether to break ties according to decision_function values (BaseSVC only)

Outputs

Name	Type	Description
support_	ndarray	Indices of support vectors
support_vectors_	ndarray	Support vectors
n_support_	ndarray	Number of support vectors for each class
dual_coef_	ndarray	Dual coefficients of support vectors in the decision function
coef_	ndarray	Weights assigned to features (linear kernel only)
intercept_	ndarray	Constants in the decision function
fit_status_	int	0 if correctly fitted, 1 otherwise
classes_	ndarray	Unique class labels (BaseSVC)
n_features_in_	int	Number of features seen during fit

Usage Examples

Basic Usage

from sklearn.svm import SVC
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split

# BaseSVC is abstract; use SVC directly
X, y = load_iris(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=42)

clf = SVC(kernel="rbf", C=1.0, gamma="scale")
clf.fit(X_train, y_train)
print(f"Accuracy: {clf.score(X_test, y_test):.3f}")
print(f"Support vectors: {clf.n_support_}")

Related Pages

• Principle:Scikit_learn_Scikit_learn_Support_Vector_Machines