nsgp (version 1.0.5)

simulategp.perturbed: Generate simulated perturbed GP models

Description

Takes pregenerated simulated GP models as input, and models both the perturbation and additional perturbation variance as GP models, that are all combined into a perturbed GP model. We sample reps data points from this at timepoints obs

Usage

simulategp.perturbed(N = 100, reps = 3, obs = c(0, 5, 10, 15, 20), gps.ctrl = NULL, filename = NULL, xs = seq(0, 20, 0.2), l.noise = 12, l.shape = 2, l.scale = 2.5, sigmaf.noise = 0.25, sigmaf.shape = 4, sigmaf.scale = 0.5)

Arguments

N

number of gp's

reps

replicate observations

obs

observation timepoints

gps.ctrl

the GPS models to be perturbed

filename

file to save the results

xs

target timepoints

l.noise

Noise model lengthscale

l.shape

shape of lengthscale Gamma distribution

l.scale

scale of lengthscale Gamma distribution

sigmaf.noise

Noise model sigma.f

sigmaf.shape

shape of sigma.f Gamma distribution

sigmaf.scale

scale of sigma.f Gamma distribution

Value

List with

simdata

Simulated datamatrix

gps

Simulated GPs

See Also

simulategp