initvars.Rd

% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/initvars.R
\name{initvars}
\alias{initvars}
\title{Initialise weights and parameters}
\usage{
initvars(
  sigma2y = 0.1,
  l_top_layer = 0.5,
  sigma2eta_top_layer = 1,
  nu = 1.5,
  transeta_mean_init = list(AWU = -3, RBF = -0.8, LFT = 1, AFF_1D = 1, AFF_2D = 1),
  transeta_mean_prior = list(AWU = -3, RBF = -0.8, LFT = NA),
  transeta_sd_init = list(AWU = 0.01, RBF = 0.01, LFT = 0.01),
  transeta_sd_prior = list(AWU = 2, RBF = 2, LFT = NA)
)
}
\arguments{
\item{sigma2y}{initial value for the measurement-error variance}

\item{l_top_layer}{initial value for the length scale at the top layer}

\item{sigma2eta_top_layer}{initial value for the variance of the weights at the top layer}

\item{nu}{initial value for the smoothness parameter}

\item{transeta_mean_init}{list of initial values for the initial weights (or the initial variational means of these weights). The list
contains five values, one for the AWU, one for the RBF, one for the LFT (Mobius), and two for the affine transformation}

\item{transeta_mean_prior}{same as \code{transeta_mean_init} but for the prior mean of the weights (SDSP only)}

\item{transeta_sd_init}{same as \code{transeta_mean_init} but for the variational standard deviations (SDSP only)}

\item{transeta_sd_prior}{same as \code{transeta_mean_init} but for the preior standard deviations of the weights (SDSP only)}
}
\value{
\code{initvars} returns a list with the initial values. Call \code{str(initvars())} to see the structure of this list.
}
\description{
Provides utility to alter the initial weights and parameters when fitting a deepspat model
}
\examples{
str(initvars(sigma2y = 0.2))
}