R: Rescale Values to and from Whole Real Number Line
rescale
R Documentation
Rescale Values to and from Whole Real Number Line
Description
Returns simulated parameters with transformations to or from the real whole number line. Parameters to transform are dependent on their specification of their prior distributions in the BUGS model.
Usage
rescale(bug, sims, to.real = NULL)
Arguments
bug
A BUGS model created in the tsbugs package.
sims
A data.frame of simulated parameter values with column names labelled according to output from the R2OpenBUGS package.
to.real
Set to TRUE or FALSE to indicate if parameters need to be scaled to the real number line or reversed scaled from the real number line back to their original parameter space.
Details
For parameters from a Gamma prior distributions in a BUGS model, a logarithmic transformation is used to convert values onto the real number line to.real=TRUE and the exponential transformation to convert vales back to their original parameter space to.real=FALSE
For parameters from Uniform and Beta prior distributions in a BUGS model, a reverse-logistic transformation is used to convert values onto the real number line to.real=TRUE and a logistic transformation to convert vales back to their original parameter space to.real=FALSE.
This function was written with the intention of reducing code when in calculating the normalising constant, in particular when 1) finding the appropiate summary statistics to simulate a candidate normalised distribution from and 2) transforming simulated values to the same scales as the orginal MCMC simulations for later use.
Value
A data.frame with the same dimensions as the sims argument.
Author(s)
Guy J. Abel
Examples
## Not run:
# demo example with constant variance models for differenced growth rate
# of England and Wales population as used in Abel et. al. (2013)
demo("cv_bma", "tsbridge")
## End(Not run)