SGDOneClassSVM#

classsklearn.linear_model.SGDOneClassSVM(nu=0.5,fit_intercept=True,max_iter=1000,tol=0.001,shuffle=True,verbose=0,random_state=None,learning_rate='optimal',eta0=0.0,power_t=0.5,warm_start=False,average=False)[source]#

Solves linear One-Class SVM using Stochastic Gradient Descent.

This implementation is meant to be used with a kernel approximationtechnique (e.g.sklearn.kernel_approximation.Nystroem) to obtain resultssimilar tosklearn.svm.OneClassSVM which uses a Gaussian kernel bydefault.

Read more in theUser Guide.

Added in version 1.0.

Parameters:

nufloat, default=0.5

The nu parameter of the One Class SVM: an upper bound on thefraction of training errors and a lower bound of the fraction ofsupport vectors. Should be in the interval (0, 1]. By default 0.5will be taken.

fit_interceptbool, default=True

Whether the intercept should be estimated or not. Defaults to True.

max_iterint, default=1000

The maximum number of passes over the training data (aka epochs).It only impacts the behavior in thefit method, and not thepartial_fit. Defaults to 1000.Values must be in the range[1,inf).

tolfloat or None, default=1e-3

The stopping criterion. If it is not None, the iterations will stopwhen (loss > previous_loss - tol). Defaults to 1e-3.Values must be in the range[0.0,inf).

shufflebool, default=True

Whether or not the training data should be shuffled after each epoch.Defaults to True.

verboseint, default=0

The verbosity level.

random_stateint, RandomState instance or None, default=None

The seed of the pseudo random number generator to use when shufflingthe data. If int, random_state is the seed used by the random numbergenerator; If RandomState instance, random_state is the random numbergenerator; If None, the random number generator is the RandomStateinstance used bynp.random.

learning_rate{‘constant’, ‘optimal’, ‘invscaling’, ‘adaptive’}, default=’optimal’

The learning rate schedule to use withfit. (If usingpartial_fit,learning rate must be controlled directly).

• ‘constant’:eta=eta0

• ‘optimal’:eta=1.0/(alpha*(t+t0))where t0 is chosen by a heuristic proposed by Leon Bottou.

• ‘invscaling’:eta=eta0/pow(t,power_t)

• ‘adaptive’: eta = eta0, as long as the training keeps decreasing.Each time n_iter_no_change consecutive epochs fail to decrease thetraining loss by tol or fail to increase validation score by tol ifearly_stopping is True, the current learning rate is divided by 5.

eta0float, default=0.0

The initial learning rate for the ‘constant’, ‘invscaling’ or‘adaptive’ schedules. The default value is 0.0 as eta0 is not used bythe default schedule ‘optimal’.Values must be in the range[0.0,inf).

power_tfloat, default=0.5

The exponent for inverse scaling learning rate.Values must be in the range(-inf,inf).

warm_startbool, default=False

When set to True, reuse the solution of the previous call to fit asinitialization, otherwise, just erase the previous solution.Seethe Glossary.

Repeatedly calling fit or partial_fit when warm_start is True canresult in a different solution than when calling fit a single timebecause of the way the data is shuffled.If a dynamic learning rate is used, the learning rate is adapteddepending on the number of samples already seen. Callingfit resetsthis counter, whilepartial_fit will result in increasing theexisting counter.

averagebool or int, default=False

When set to True, computes the averaged SGD weights and stores theresult in thecoef_ attribute. If set to an int greater than 1,averaging will begin once the total number of samples seen reachesaverage. Soaverage=10 will begin averaging after seeing 10samples.

Attributes:

coef_ndarray of shape (1, n_features)

Weights assigned to the features.

offset_ndarray of shape (1,)

Offset used to define the decision function from the raw scores.We have the relation: decision_function = score_samples - offset.

n_iter_int

The actual number of iterations to reach the stopping criterion.

t_int

Number of weight updates performed during training.Same as(n_iter_*n_samples+1).

n_features_in_int

Number of features seen duringfit.

Added in version 0.24.

feature_names_in_ndarray of shape (n_features_in_,)

Names of features seen duringfit. Defined only whenXhas feature names that are all strings.

Added in version 1.0.

See also

sklearn.svm.OneClassSVM

Unsupervised Outlier Detection.

Notes

This estimator has a linear complexity in the number of training samplesand is thus better suited than thesklearn.svm.OneClassSVMimplementation for datasets with a large number of training samples (say> 10,000).

Examples

>>>importnumpyasnp>>>fromsklearnimportlinear_model>>>X=np.array([[-1,-1],[-2,-1],[1,1],[2,1]])>>>clf=linear_model.SGDOneClassSVM(random_state=42)>>>clf.fit(X)SGDOneClassSVM(random_state=42)

>>>print(clf.predict([[4,4]]))[1]

decision_function(X)[source]#

Signed distance to the separating hyperplane.

Signed distance is positive for an inlier and negative for anoutlier.

Parameters:

X{array-like, sparse matrix}, shape (n_samples, n_features)

Testing data.

Returns:

decarray-like, shape (n_samples,)

Decision function values of the samples.

densify()[source]#

Convert coefficient matrix to dense array format.

Converts thecoef_ member (back) to a numpy.ndarray. This is thedefault format ofcoef_ and is required for fitting, so callingthis method is only required on models that have previously beensparsified; otherwise, it is a no-op.

Returns:

self

Fitted estimator.

fit(X,y=None,coef_init=None,offset_init=None,sample_weight=None)[source]#

Fit linear One-Class SVM with Stochastic Gradient Descent.

This solves an equivalent optimization problem of theOne-Class SVM primal optimization problem and returns a weight vectorw and an offset rho such that the decision function is given by<w, x> - rho.

Parameters:

X{array-like, sparse matrix}, shape (n_samples, n_features)

Training data.

yIgnored

Not used, present for API consistency by convention.

coef_initarray, shape (n_classes, n_features)

The initial coefficients to warm-start the optimization.

offset_initarray, shape (n_classes,)

The initial offset to warm-start the optimization.

sample_weightarray-like, shape (n_samples,), optional

Weights applied to individual samples.If not provided, uniform weights are assumed. These weights willbe multiplied with class_weight (passed through theconstructor) if class_weight is specified.

Returns:

selfobject

Returns a fitted instance of self.

fit_predict(X,y=None,**kwargs)[source]#

Perform fit on X and returns labels for X.

Returns -1 for outliers and 1 for inliers.

Parameters:

X{array-like, sparse matrix} of shape (n_samples, n_features)

The input samples.

yIgnored

Not used, present for API consistency by convention.

**kwargsdict

Arguments to be passed tofit.

Added in version 1.4.

Returns:

yndarray of shape (n_samples,)

1 for inliers, -1 for outliers.

get_metadata_routing()[source]#

Get metadata routing of this object.

Please checkUser Guide on how the routingmechanism works.

Returns:

routingMetadataRequest

AMetadataRequest encapsulatingrouting information.

get_params(deep=True)[source]#

Get parameters for this estimator.

Parameters:

deepbool, default=True

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

Returns:

paramsdict

Parameter names mapped to their values.

partial_fit(X,y=None,sample_weight=None)[source]#

Fit linear One-Class SVM with Stochastic Gradient Descent.

Parameters:

X{array-like, sparse matrix}, shape (n_samples, n_features)

Subset of the training data.

yIgnored

Not used, present for API consistency by convention.

sample_weightarray-like, shape (n_samples,), optional

Weights applied to individual samples.If not provided, uniform weights are assumed.

Returns:

selfobject

Returns a fitted instance of self.

predict(X)[source]#

Return labels (1 inlier, -1 outlier) of the samples.

Parameters:

X{array-like, sparse matrix}, shape (n_samples, n_features)

Testing data.

Returns:

yarray, shape (n_samples,)

Labels of the samples.

score_samples(X)[source]#

Raw scoring function of the samples.

Parameters:

X{array-like, sparse matrix}, shape (n_samples, n_features)

Testing data.

Returns:

score_samplesarray-like, shape (n_samples,)

Unshiffted scoring function values of the samples.

set_fit_request(*,coef_init:bool|None|str='$UNCHANGED$',offset_init:bool|None|str='$UNCHANGED$',sample_weight:bool|None|str='$UNCHANGED$')→SGDOneClassSVM[source]#

Configure whether metadata should be requested to be passed to thefit method.

Note that this method is only relevant when this estimator is used as asub-estimator within ameta-estimator and metadata routing is enabledwithenable_metadata_routing=True (seesklearn.set_config).Please check theUser Guide on how the routingmechanism works.

The options for each parameter are:

• True: metadata is requested, and passed tofit if provided. The request is ignored if metadata is not provided.

• False: metadata is not requested and the meta-estimator will not pass it tofit.

• 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.

Parameters:

coef_initstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing forcoef_init parameter infit.

offset_initstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing foroffset_init parameter infit.

sample_weightstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing forsample_weight parameter infit.

Returns:

selfobject

The updated object.

set_params(**params)[source]#

Set the parameters of this estimator.

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

Parameters:

**paramsdict

Estimator parameters.

Returns:

selfestimator instance

Estimator instance.

set_partial_fit_request(*,sample_weight:bool|None|str='$UNCHANGED$')→SGDOneClassSVM[source]#

Configure whether metadata should be requested to be passed to thepartial_fit method.

Note that this method is only relevant when this estimator is used as asub-estimator within ameta-estimator and metadata routing is enabledwithenable_metadata_routing=True (seesklearn.set_config).Please check theUser Guide on how the routingmechanism works.

The options for each parameter are:

• True: metadata is requested, and passed topartial_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 topartial_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.

Parameters:

sample_weightstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing forsample_weight parameter inpartial_fit.

Returns:

selfobject

The updated object.

sparsify()[source]#

Convert coefficient matrix to sparse format.

Converts thecoef_ member to a scipy.sparse matrix, which forL1-regularized models can be much more memory- and storage-efficientthan the usual numpy.ndarray representation.

Theintercept_ member is not converted.

Returns:

self

Fitted estimator.

Notes

For non-sparse models, i.e. when there are not many zeros incoef_,this may actuallyincrease memory usage, so use this method withcare. A rule of thumb is that the number of zero elements, which canbe computed with(coef_==0).sum(), must be more than 50% for thisto provide significant benefits.

After calling this method, further fitting with the partial_fitmethod (if any) will not work until you call densify.

Gallery examples#

One-Class SVM versus One-Class SVM using Stochastic Gradient Descent

One-Class SVM versus One-Class SVM using Stochastic Gradient Descent

Comparing anomaly detection algorithms for outlier detection on toy datasets

Comparing anomaly detection algorithms for outlier detection on toy datasets