himalaya.kernel_ridge.KernelRidge

class himalaya.kernel_ridge.KernelRidge(alpha=1, kernel='linear', kernel_params=None, solver='auto', solver_params=None, fit_intercept=False, force_cpu=False, warn=True)

Kernel ridge regression.

Solve the kernel ridge regression:

w* = argmin_w ||K @ w - Y||^2 + alpha (w.T @ K @ w)

where K is a kernel computed on the input X. The default kernel is linear (K = X @ X.T).

Parameters

alphafloat, or array of shape (n_targets, )

L2 regularization parameter.

kernelstr or callable, default=”linear”

Kernel mapping. Available kernels are: ‘linear’, ‘polynomial, ‘poly’, ‘rbf’, ‘sigmoid’, ‘cosine’, or ‘precomputed’. Set to ‘precomputed’ in order to pass a precomputed kernel matrix to the estimator methods instead of samples. A callable should accept two arguments and the keyword arguments passed to this object as kernel_params, and should return a floating point number.

kernel_paramsdict or None

Additional parameters for the kernel function. See more details in the docstring of the function: KernelRidge.ALL_KERNELS[kernel]

solverstr

Algorithm used during the fit, “eigenvalues”, “conjugate_gradient”, “gradient_descent”, or “auto”. If “auto”, use “eigenvalues” if alpha is a float, and “conjugate_gradient” is alpha is an array.

solver_paramsdict or None

Additional parameters for the solver. See more details in the docstring of the function: KernelRidge.ALL_SOLVERS[solver]

fit_interceptboolean

Whether to fit an intercept. If False, X and Y must be zero-mean over samples.

force_cpubool

If True, computations will be performed on CPU, ignoring the current backend. If False, use the current backend.

warnbool

If True, warn if the number of samples is larger than the number of features, and if the kernel is linear.

Examples

>>> from himalaya.kernel_ridge import KernelRidge
>>> import numpy as np
>>> n_samples, n_features, n_targets = 10, 5, 3
>>> X = np.random.randn(n_samples, n_features)
>>> Y = np.random.randn(n_samples, n_targets)
>>> model = KernelRidge()
>>> model.fit(X, Y)
KernelRidge()

Attributes

dual_coef_array of shape (n_samples) or (n_samples, n_targets)

Representation of weight vectors in kernel space.

intercept_float or array of shape (n_targets, )

Intercept. Only present if fit_intercept is True.

X_fit_array of shape (n_samples, n_features)

Training data. If kernel == “precomputed” this is None.

n_features_in_int

Number of features (or number of samples if kernel == “precomputed”) used during the fit.

dtype_str

Dtype of input data.

Methods

fit(X[, y, sample_weight])	Fit the model.
get_params([deep])	Get parameters for this estimator.
get_primal_coef([X_fit])	Returns the primal coefficients, assuming the kernel is linear.
predict(X)	Predict using the model.
score(X, y)	Return the coefficient of determination R^2 of the prediction.
set_params(**params)	Set the parameters of this estimator.