inst/doc/picking_priors.R

In fido: Bayesian Multinomial Logistic Normal Regression

## ----include = FALSE----------------------------------------------------------
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

## ----eval=FALSE---------------------------------------------------------------
#  # To put prior on ALR_j coordinates for some j in (1,...,D-1)
#  Xi <- clrvar2alrvar(S, j)
#  # To put prior in a particular ILR coordinate defined by contrast matrix V
#  Xi <- clrvar2ilrvar(S, V)
#  # To put prior in CLR coordinates (this one needs two transforms)
#  foo <- clrvar2alrvar(S, D)
#  Xi <- alrvar2clrvar(foo, D)

## ----eval=FALSE---------------------------------------------------------------
#  # Transform from log-absolute-abundance effects to effects on absolute-abundances
#  foo <- exp(A)
#  # To put prior on ALR_j coordinates for some j in (1,...,D-1)
#  Theta <- driver::alr_array(foo, j, parts=1)
#  # To put prior in a particular ILR coordinate defined by contrast matrix V
#  Theta <- driver::ilr_array(foo, V, parts=1)
#  # To put prior in CLR coordinates
#  Theta <- driver::clr_array(foo, parts=1)