param: Get or set GP model parameters

In gplite: General Purpose Gaussian Process Modelling

Get or set GP model parameters

Description

get_param returns the current hyperparameters of the GP model in a vector. set_param can be used to set the parameters. Note that these functions are intended mainly for internal usage, and there is typically no need to use these functions directly but instead create a new GP model using gp_init.

Usage

get_param(object, ...)

set_param(object, param, ...)

Arguments

object	The model object.
...	Ignored currently.
param	The parameters to be set. Call first get_param to see the order in which the parameters should be given for a particular model. Notice that all positive parameters should be given in a log-scale.

Value

get_param returns the current hyperparameters and set_param the GP model structure with the new parameter values.

Examples

# Set up some model
gp <- gp_init(cf = cf_sexp(), lik = lik_gaussian())

# print out to see the parameter ordering
param <- get_param(gp)
print(param)

# set some new values
param_new <- log(c(0.1, 0.8, 0.3))
names(param_new) <- names(param)
gp <- set_param(gp, param_new)

# check the result
print(get_param(gp))

gplite documentation built on Aug. 24, 2022, 9:07 a.m.