Gaussian Process Based Design and Analysis for the Active Subspace Method

[Up] [Top]

Documentation for package ‘activegp’ version 1.1.1

Help Pages

activegp Package activegp
as.matrix.const_C Extract Matrix
C_GP Active Subspace Matrix closed form expression for a GP.
C_GP_ci CI on Eigenvalues via Monte Carlo/GP
C_tr Expected variance of trace of C
C_var Element-wise Cn+1 variance
C_var2 Alternative Variance of Update
domain_to_R Rectangular Domain -> Unbounded Domain
domain_to_unit Change a function's inputs to live in [-1, 1]
grad_est_subspace Estimate the Active Subspace of a Cheap Function using Gradients
logLikHessian Hessian of the log-likelihood with respect to lengthscales hyperparameters Works for homGP and hetGP models from the hetGP package for now.
Lt_GP Active Subspace Prewarping
n11_2_01 f:[-1, 1] -> R Becomes f:[0,1] -> R
plot.const_C Plot const_C objectc
print.const_C Print const_C objects
subspace_dist Get the distance between subspaces defined as the ranges of A and B
update.const_C C update with new observations
update_C2 Update Constantine's C, using update formula