parameter_VAR: Varying parameter object constructor.

In BaM-tools/RBaM: Bayesian Modeling: Estimate a Computer Model and Make Uncertain Predictions

Varying parameter object constructor.

Description

Creates a new instance of a 'parameter_VAR' object

Usage

parameter_VAR(
  name,
  index,
  d,
  init,
  prior.dist = rep("FlatPrior", length(init)),
  prior.par = rep(list(NULL), length(init))
)

Arguments

name	character, parameter name.
index	character, name of column in VAR.indx (see ?dataset) containing the index for this varying parameter
d	dataset object, the dataset containing (amongst other things) the index above
init	numeric vector, initial guesses for each instance of the VAR parameter.
prior.dist	character vector, prior distribution for each instance of the VAR parameter.
prior.par	list of numeric vectors, prior parameters for each instance of the VAR parameter

Value

An object of class 'parameter_VAR'.

Examples

X=data.frame(input1=rnorm(100),input2=rnorm(100))
Y=data.frame(output=X$input1+0.8*X$input2+0.1*rnorm(100))
VAR.indx=data.frame(indx=c(rep(1,50),rep(2,50)))
workspace=tempdir()
d <- dataset(X=X,Y=Y,data.dir=workspace,VAR.indx=VAR.indx)
p <- parameter_VAR(name='par',index='indx',d=d,
                   init=c(-1,1,2),
                   prior.dist=c('Gaussian','FlatPrior','Triangle'),
                   prior.par=list(c(-1,1),NULL,c(2,0,5)))

BaM-tools/RBaM documentation built on April 11, 2025, 10:01 p.m.