sgp: Fit Sparse Gaussian Process via variational inference

In DongyueXie/smashrgen: Empirical Bayes smoothers

Fit Sparse Gaussian Process via variational inference

Description

Fit Sparse Gaussian Process via variational inference

Usage

sgp(
  y,
  X = NULL,
  X_ind = NULL,
  m = 10,
  kernel_func = Maternkernel,
  kernel_param = c(0, 0),
  mu = NULL,
  s2 = NULL,
  sigma2 = NULL,
  opt_method = "L-BFGS-B",
  fix_X_ind = T,
  fix_kernel_param = F,
  fix_sigma2 = F,
  fix_mu = F,
  l_b = -Inf,
  r_b = Inf,
  Jitter = 1e-05,
  n_restart = 5,
  verbose = FALSE
)

Arguments

X, y	training data X and response y.
X_ind	inducing point locations.
m	default number of X_ind is 10, if X_ind is not given as an input
kernel_func, kernel_param	Kernel functions to use, and their parameters
mu	prior mean
s2	known variances
sigma2	noise variance
opt_method	optimization method for estimating prior parameters in 'optim'
l_b, r_b	lower and upper bound of prior parameters
Jitter	added to diagonal of the Kernel matrix for numerical stability
n_restart	number of re-start of different kernel params values
fix_	whether fix those parameters

DongyueXie/smashrgen documentation built on Jan. 14, 2024, 5:30 a.m.