Multi Kernels

Generate multiple kernel views from a single feature matrix.

class polyview.augmentation.multi_kernels.MultiKernel(*args: Any, **kwargs: Any)

Bases: BaseEstimator

Generate multiple kernel-matrix views from one input matrix.

Parameters:

• specs (sequence of KernelSpec or kernel-name strings, optional) – Kernel definitions to apply to the same input matrix. If omitted, defaults to three complementary kernels: linear, rbf, and polynomial.

• view_names (sequence of str, optional) – Names to assign to the generated views.

fit(X, y=None) → MultiKernel

fit_transform(X, y=None) → MultiViewDataset

transform(X) → MultiViewDataset

polyview.augmentation.multi_kernels.multi_kernels(X, specs: Sequence[KernelSpec | Literal['linear', 'rbf', 'polynomial', 'precomputed']] | None = None, labels=None, view_names: Sequence[str] | None = None) → MultiViewDataset

Build a MultiViewDataset of kernel matrices from a single matrix.