API reference

CompositeKernel

class svdynamics.CompositeKernel(kernels: Sequence[str | Tuple[str, Dict[str, Any]] | Any], weights: Sequence[float] | None = None, normalize: bool = True, eps: float = 1e-12)

Bases: object

CompositeKernel builds a callable kernel that is a weighted sum of base kernels.

K(X, Y) = sum_m w_m * K_m(X, Y)

Parameters

kernels:

List of kernel specs. Each element can be: - “rbf” | “linear” | “poly” | “sigmoid” - (“rbf”, {“gamma”: 0.1}) etc. - callable kernel(X, Y) -> ndarray

weights:

Optional weights for kernels. If None, uniform weights are used. Weights are normalized to sum to 1.

normalize:

If True, each base kernel is rescaled by a scalar derived from K(X, X) to reduce dominance due to scale.

eps:

Small constant used for numerical stability.

as_callable() → Callable[[ndarray, ndarray], ndarray]

eps: float = 1e-12

kernels: Sequence[str | Tuple[str, Dict[str, Any]] | Any]

normalize: bool = True

property normalized_weights: ndarray

property parsed_kernels: List[ParsedKernel]

weights: Sequence[float] | None = None

SVDClassifier

class svdynamics.SVDClassifier(C: float = 1.0, kernel: str | CompositeKernel | Any = 'rbf', degree: int = 3, gamma: str | float = 'scale', coef0: float = 0.0, shrinking: bool = True, probability: bool = False, tol: float = 0.001, cache_size: float = 200.0, class_weight: Dict[Any, float] | None = None, verbose: bool = False, max_iter: int = -1, decision_function_shape: str = 'ovr', break_ties: bool = False, random_state: int | None = None)

Bases: BaseEstimator, ClassifierMixin

SVDClassifier is a thin wrapper around sklearn.svm.SVC that supports CompositeKernel.

If kernel is a CompositeKernel, the estimator uses a callable kernel. Otherwise it passes through to sklearn SVC.

C: float = 1.0

break_ties: bool = False

cache_size: float = 200.0

class_weight: Dict[Any, float] | None = None

coef0: float = 0.0

decision_function(X: ndarray) → ndarray

decision_function_shape: str = 'ovr'

degree: int = 3

fit(X: ndarray, y: ndarray) → SVDClassifier

gamma: str | float = 'scale'

kernel: str | CompositeKernel | Any = 'rbf'

max_iter: int = -1

property n_support_: ndarray

predict(X: ndarray) → ndarray

predict_proba(X: ndarray) → ndarray

probability: bool = False

random_state: int | None = None

set_score_request(*, sample_weight: bool | None | str = '$UNCHANGED$') → SVDClassifier

Request metadata passed to the score method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism 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 the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Note

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

Parameters

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

Metadata routing for sample_weight parameter in score.

Returns

selfobject

The updated object.

shrinking: bool = True

property support_vectors_: ndarray

tol: float = 0.001

verbose: bool = False

SVDRegressor

class svdynamics.SVDRegressor(C: float = 1.0, kernel: str | CompositeKernel | Any = 'rbf', degree: int = 3, gamma: str | float = 'scale', coef0: float = 0.0, tol: float = 0.001, epsilon: float = 0.1, shrinking: bool = True, cache_size: float = 200.0, verbose: bool = False, max_iter: int = -1)

Bases: BaseEstimator, RegressorMixin

SVDRegressor is a thin wrapper around sklearn.svm.SVR that supports CompositeKernel.

C: float = 1.0

cache_size: float = 200.0

coef0: float = 0.0

degree: int = 3

epsilon: float = 0.1

fit(X: ndarray, y: ndarray) → SVDRegressor

gamma: str | float = 'scale'

kernel: str | CompositeKernel | Any = 'rbf'

max_iter: int = -1

predict(X: ndarray) → ndarray

set_score_request(*, sample_weight: bool | None | str = '$UNCHANGED$') → SVDRegressor

Request metadata passed to the score method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism 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 the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Note

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

Parameters

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

Metadata routing for sample_weight parameter in score.

Returns

selfobject

The updated object.

shrinking: bool = True

property support_: ndarray

tol: float = 0.001

verbose: bool = False