rgpycrumbs.surfaces.gradient

Classes

GradientMatern	Gradient-enhanced Matérn 5/2 surface implementation.
GradientSE	Gradient-enhanced Squared Exponential (SE) surface implementation.
GradientIMQ	Gradient-enhanced Inverse Multi-Quadratic (IMQ) surface implementation.
GradientRQ	Symmetric Gradient-enhanced Rational Quadratic (RQ) surface implementation.
NystromGradientIMQ	Memory-efficient Nystrom-approximated gradient-enhanced IMQ surface.

Functions

negative_mll_matern_grad(log_params, x, y_flat, D_plus_1)
_grad_matern_solve(x, y_full, noise_scalar, length_scale)
_grad_matern_predict(x_query, x_obs, alpha, length_scale)
_grad_matern_var(x_query, x_obs, K_inv, length_scale)
negative_mll_se_grad(log_params, x, y_flat, D_plus_1)
_grad_se_solve(x, y_full, noise_scalar, length_scale)
_grad_se_predict(x_query, x_obs, alpha, length_scale)
_grad_se_var(x_query, x_obs, K_inv, length_scale)
negative_mll_imq_grad(log_params, x, y_flat, D_plus_1)
negative_mll_imq_map(log_params, init_eps, x, y_flat, ...)
_grad_imq_solve(x, y_full, noise_scalar, epsilon)
_grad_imq_predict(x_query, x_obs, alpha, epsilon)
_grad_imq_var(x_query, x_obs, K_inv, epsilon)
negative_mll_rq_map(log_params, x, y_flat, D_plus_1)
_grad_rq_solve(x, y_full, noise_scalar, params)
_grad_rq_predict(x_query, x_obs, alpha, params)
_grad_rq_var(x_query, x_obs, K_inv, params)
_stable_nystrom_grad_imq_solve(x, y_full, x_inducing, ...)
_nystrom_grad_imq_predict(x_query, x_inducing, ...)
_nystrom_grad_imq_var(x_query, x_inducing, W, epsilon)

Module Contents

rgpycrumbs.surfaces.gradient.negative_mll_matern_grad(log_params, x, y_flat, D_plus_1)

rgpycrumbs.surfaces.gradient._grad_matern_solve(x, y_full, noise_scalar, length_scale)

rgpycrumbs.surfaces.gradient._grad_matern_predict(x_query, x_obs, alpha, length_scale)

rgpycrumbs.surfaces.gradient._grad_matern_var(x_query, x_obs, K_inv, length_scale)

class rgpycrumbs.surfaces.gradient.GradientMatern(x, y, gradients=None, smoothing=0.0001, length_scale=None, optimize=True, **_kwargs)

Bases: rgpycrumbs.surfaces._base.BaseGradientSurface

Gradient-enhanced Matérn 5/2 surface implementation.

_fit(smoothing, length_scale, optimize)

Internal method to perform parameter optimization.

_solve()

Internal method to solve the linear system for weights.

_predict_chunk(chunk)

Internal method for batch prediction.

_var_chunk(chunk)

Internal method for batch variance.

rgpycrumbs.surfaces.gradient.negative_mll_se_grad(log_params, x, y_flat, D_plus_1)

rgpycrumbs.surfaces.gradient._grad_se_solve(x, y_full, noise_scalar, length_scale)

rgpycrumbs.surfaces.gradient._grad_se_predict(x_query, x_obs, alpha, length_scale)

rgpycrumbs.surfaces.gradient._grad_se_var(x_query, x_obs, K_inv, length_scale)

class rgpycrumbs.surfaces.gradient.GradientSE(x, y, gradients=None, smoothing=0.0001, length_scale=None, optimize=True, **_kwargs)

Bases: rgpycrumbs.surfaces._base.BaseGradientSurface

Gradient-enhanced Squared Exponential (SE) surface implementation.

_fit(smoothing, length_scale, optimize)

Internal method to perform parameter optimization.

_solve()

Internal method to solve the linear system for weights.

_predict_chunk(chunk)

Internal method for batch prediction.

_var_chunk(chunk)

Internal method for batch variance.

rgpycrumbs.surfaces.gradient.negative_mll_imq_grad(log_params, x, y_flat, D_plus_1)

rgpycrumbs.surfaces.gradient.negative_mll_imq_map(log_params, init_eps, x, y_flat, D_plus_1)

rgpycrumbs.surfaces.gradient._grad_imq_solve(x, y_full, noise_scalar, epsilon)

rgpycrumbs.surfaces.gradient._grad_imq_predict(x_query, x_obs, alpha, epsilon)

rgpycrumbs.surfaces.gradient._grad_imq_var(x_query, x_obs, K_inv, epsilon)

class rgpycrumbs.surfaces.gradient.GradientIMQ(x, y, gradients=None, smoothing=0.0001, length_scale=None, optimize=True, **_kwargs)

Bases: rgpycrumbs.surfaces._base.BaseGradientSurface

Gradient-enhanced Inverse Multi-Quadratic (IMQ) surface implementation.

_fit(smoothing, length_scale, optimize)

Internal method to perform parameter optimization.

_solve()

Internal method to solve the linear system for weights.

_predict_chunk(chunk)

Internal method for batch prediction.

_var_chunk(chunk)

Internal method for batch variance.

rgpycrumbs.surfaces.gradient.negative_mll_rq_map(log_params, x, y_flat, D_plus_1)

rgpycrumbs.surfaces.gradient._grad_rq_solve(x, y_full, noise_scalar, params)

rgpycrumbs.surfaces.gradient._grad_rq_predict(x_query, x_obs, alpha, params)

rgpycrumbs.surfaces.gradient._grad_rq_var(x_query, x_obs, K_inv, params)

class rgpycrumbs.surfaces.gradient.GradientRQ(x, y, gradients=None, smoothing=0.0001, length_scale=None, optimize=True, **_kwargs)

Bases: rgpycrumbs.surfaces._base.BaseGradientSurface

Symmetric Gradient-enhanced Rational Quadratic (RQ) surface implementation.

_fit(_smoothing, length_scale, optimize)

Internal method to perform parameter optimization.

_solve()

Internal method to solve the linear system for weights.

_predict_chunk(chunk)

Internal method for batch prediction.

_var_chunk(chunk)

Internal method for batch variance.

rgpycrumbs.surfaces.gradient._stable_nystrom_grad_imq_solve(x, y_full, x_inducing, noise_scalar, epsilon)

rgpycrumbs.surfaces.gradient._nystrom_grad_imq_predict(x_query, x_inducing, alpha_m, epsilon)

rgpycrumbs.surfaces.gradient._nystrom_grad_imq_var(x_query, x_inducing, W, epsilon)

class rgpycrumbs.surfaces.gradient.NystromGradientIMQ(x, y, gradients=None, n_inducing=300, nimags=None, smoothing=0.001, length_scale=None, optimize=True, **kwargs)

Bases: rgpycrumbs.surfaces._base.BaseGradientSurface

Memory-efficient Nystrom-approximated gradient-enhanced IMQ surface.

n_inducing = 300

nimags = None

_fit(smoothing, length_scale, _optimize)

Internal method to perform parameter optimization.

_solve()

Internal method to solve the linear system for weights.

_predict_chunk(chunk)

Internal method for batch prediction.

_var_chunk(chunk)

Internal method for batch variance.