C-Support Vector Classification.

The implementation is based on libsvm. The fit time scales at leastquadratically with the number of samples and may be impracticalbeyond tens of thousands of samples. For large datasetsconsider using LinearSVC orSGDClassifier instead, possibly after aNystroem transformer orother Kernel Approximation.

The multiclass support is handled according to a one-vs-one scheme.

For details on the precise mathematical formulation of the providedkernel functions and how gamma, coef0 and degree affect eachother, see the corresponding section in the narrative documentation:Kernel functions.

To learn how to tune SVC’s hyperparameters, see the following example:Nested versus non-nested cross-validation

Read more in the User Guide.

Regularization parameter. The strength of the regularization isinversely proportional to C. Must be strictly positive. The penaltyis a squared l2 penalty. For an intuitive visualization of the effectsof scaling the regularization parameter C, seeScaling the regularization parameter for SVCs.

Specifies the kernel type to be used in the algorithm. Ifnone is given, ‘rbf’ will be used. If a callable is given it is used topre-compute the kernel matrix from data matrices; that matrix should bean array of shape (n_samples, n_samples). For an intuitivevisualization of different kernel types seePlot classification boundaries with different SVM Kernels.

Degree of the polynomial kernel function (‘poly’).Must be non-negative. Ignored by all other kernels.

Kernel coefficient for ‘rbf’, ‘poly’ and ‘sigmoid’.

if gamma='scale' (default) is passed then it uses1 / (n_features * X.var()) as value of gamma,

if ‘auto’, uses 1 / n_features

if float, must be non-negative.

Changed in version 0.22: The default value of gamma changed from ‘auto’ to ‘scale’.

Independent term in kernel function.It is only significant in ‘poly’ and ‘sigmoid’.

Whether to use the shrinking heuristic.See the User Guide.

Whether to enable probability estimates. This must be enabled priorto calling fit, will slow down that method as it internally uses5-fold cross-validation, and predict_proba may be inconsistent withpredict. Read more in the User Guide.

Tolerance for stopping criterion.

Specify the size of the kernel cache (in MB).

Set the parameter C of class i to class_weight[i]*C forSVC. If not given, all classes are supposed to haveweight one.The “balanced” mode uses the values of y to automatically adjustweights inversely proportional to class frequencies in the input dataas n_samples / (n_classes * np.bincount(y)).

Enable verbose output. Note that this setting takes advantage of aper-process runtime setting in libsvm that, if enabled, may not workproperly in a multithreaded context.

Hard limit on iterations within solver, or -1 for no limit.

Whether to return a one-vs-rest (‘ovr’) decision function of shape(n_samples, n_classes) as all other classifiers, or the originalone-vs-one (‘ovo’) decision function of libsvm which has shape(n_samples, n_classes * (n_classes - 1) / 2). However, note thatinternally, one-vs-one (‘ovo’) is always used as a multi-class strategyto train models; an ovr matrix is only constructed from the ovo matrix.The parameter is ignored for binary classification.

Changed in version 0.19: decision_function_shape is ‘ovr’ by default.

Added in version 0.17: decision_function_shape=’ovr’ is recommended.

Changed in version 0.17: Deprecated decision_function_shape=’ovo’ and None.

If true, decision_function_shape='ovr', and number of classes > 2,predict will break ties according to the confidence values ofdecision_function; otherwise the first class among the tiedclasses is returned. Please note that breaking ties comes at arelatively high computational cost compared to a simple predict.

Added in version 0.22.

Controls the pseudo random number generation for shuffling the data forprobability estimates. Ignored when probability is False.Pass an int for reproducible output across multiple function calls.See Glossary.

Multipliers of parameter C for each class.Computed based on the class_weight parameter.

The classes labels.

Weights assigned to the features when kernel="linear".

Dual coefficients of the support vector in the decisionfunction (see Mathematical formulation), multiplied bytheir targets.For multiclass, coefficient for all 1-vs-1 classifiers.The layout of the coefficients in the multiclass case is somewhatnon-trivial. See the multi-class section of the User Guide for details.

0 if correctly fitted, 1 otherwise (will raise warning)

Constants in decision function.

Number of features seen during fit.

Added in version 0.24.

Names of features seen during fit. Defined only when Xhas feature names that are all strings.

Added in version 1.0.

Number of iterations run by the optimization routine to fit the model.The shape of this attribute depends on the number of models optimizedwhich in turn depends on the number of classes.

Added in version 1.1.

Indices of support vectors.

Support vectors. An empty array if kernel is precomputed.

Number of support vectors for each class.

Parameter learned in Platt scaling when probability=True.

Parameter learned in Platt scaling when probability=True.

Array dimensions of training vector X.

See also

Support Vector Machine for Regression implemented using libsvm.

Scalable Linear Support Vector Machine for classification implemented using liblinear. Check the See Also section of LinearSVC for more comparison element.

References

LIBSVM: A Library for Support Vector Machines

Platt, John (1999). “Probabilistic Outputs for Support VectorMachines and Comparisons to Regularized Likelihood Methods”

Examples

Weights assigned to the features when kernel="linear".

Evaluate the decision function for the samples in X.

The input samples.

Returns the decision function of the sample for each classin the model.If decision_function_shape=’ovr’, the shape is (n_samples,n_classes).

Notes

If decision_function_shape=’ovo’, the function values are proportionalto the distance of the samples X to the separating hyperplane. If theexact distances are required, divide the function values by the norm ofthe weight vector (coef_). See also this question for further details.If decision_function_shape=’ovr’, the decision function is a monotonictransformation of ovo decision function.

Fit the SVM model according to the given training data.

Training vectors, where n_samples is the number of samplesand n_features is the number of features.For kernel=”precomputed”, the expected shape of X is(n_samples, n_samples).

Target values (class labels in classification, real numbers inregression).

Per-sample weights. Rescale C per sample. Higher weightsforce the classifier to put more emphasis on these points.

Fitted estimator.

Notes

If X and y are not C-ordered and contiguous arrays of np.float64 andX is not a scipy.sparse.csr_matrix, X and/or y may be copied.

If X is a dense array, then the other methods will not support sparsematrices as input.

Get metadata routing of this object.

Please check User Guide on how the routingmechanism works.

A MetadataRequest encapsulatingrouting information.

Get parameters for this estimator.

If True, will return the parameters for this estimator andcontained subobjects that are estimators.

Parameter names mapped to their values.

Number of support vectors for each class.

Perform classification on samples in X.

For an one-class model, +1 or -1 is returned.

For kernel=”precomputed”, the expected shape of X is(n_samples_test, n_samples_train).

Class labels for samples in X.

Compute log probabilities of possible outcomes for samples in X.

The model need to have probability information computed at trainingtime: fit with attribute probability set to True.

For kernel=”precomputed”, the expected shape of X is(n_samples_test, n_samples_train).

Returns the log-probabilities of the sample for each class inthe model. The columns correspond to the classes in sortedorder, as they appear in the attribute classes_.

Notes

The probability model is created using cross validation, sothe results can be slightly different than those obtained bypredict. Also, it will produce meaningless results on very smalldatasets.

Compute probabilities of possible outcomes for samples in X.

The model needs to have probability information computed at trainingtime: fit with attribute probability set to True.

For kernel=”precomputed”, the expected shape of X is(n_samples_test, n_samples_train).

Returns the probability of the sample for each class inthe model. The columns correspond to the classes in sortedorder, as they appear in the attribute classes_.

Notes

The probability model is created using cross validation, sothe results can be slightly different than those obtained bypredict. Also, it will produce meaningless results on very smalldatasets.

Parameter learned in Platt scaling when probability=True.

Parameter learned in Platt scaling when probability=True.

Return the mean accuracy on the given test data and labels.

In multi-label classification, this is the subset accuracywhich is a harsh metric since you require for each sample thateach label set be correctly predicted.

Test samples.

True labels for X.

Sample weights.

Mean accuracy of self.predict(X) w.r.t. y.

Request metadata passed to the fit method.

Note that this method is only relevant ifenable_metadata_routing=True (see sklearn.set_config).Please see User Guide on how the routingmechanism works.

The options for each parameter are:

True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

False: metadata is not requested and the meta-estimator will not pass it to fit.

None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains theexisting request. This allows you to change the request for someparameters and not others.

Added in version 1.3.

Note

This method is only relevant if this estimator is used as asub-estimator of a meta-estimator, e.g. used inside aPipeline. Otherwise it has no effect.

Metadata routing for sample_weight parameter in fit.

The updated object.

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects(such as Pipeline). The latter haveparameters of the form __ so that it’spossible to update each component of a nested object.

Estimator parameters.

Estimator instance.

Request metadata passed to the score method.

Note that this method is only relevant ifenable_metadata_routing=True (see sklearn.set_config).Please see User Guide on how the routingmechanism works.

The options for each parameter are:

True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

False: metadata is not requested and the meta-estimator will not pass it to score.

None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains theexisting request. This allows you to change the request for someparameters and not others.

Added in version 1.3.

Note

This method is only relevant if this estimator is used as asub-estimator of a meta-estimator, e.g. used inside aPipeline. Otherwise it has no effect.

Metadata routing for sample_weight parameter in score.

The updated object.