From 4b43ac83d0bcc581a8dfb418471cf869e9ae964b Mon Sep 17 00:00:00 2001 From: fweber144 Date: Fri, 3 Nov 2023 10:27:28 +0100 Subject: [PATCH] Docs: Enhance the `nloo` entry. --- R/cv_varsel.R | 10 +++++----- man/cv_varsel.Rd | 10 +++++----- 2 files changed, 10 insertions(+), 10 deletions(-) diff --git a/R/cv_varsel.R b/R/cv_varsel.R index 1bcc30d51..628d9174e 100644 --- a/R/cv_varsel.R +++ b/R/cv_varsel.R @@ -22,11 +22,11 @@ #' performed. See also section "Note" below. #' @param nloo **Caution:** Still experimental. Only relevant if `cv_method = #' "LOO"`. Number of subsampled PSIS-LOO CV folds, i.e., number of -#' observations used for the approximate LOO CV (anything between 1 and the -#' original number of observations). Smaller values lead to faster computation -#' but higher uncertainty in the evaluation part. If `NULL`, all observations -#' are used, but for faster experimentation, one can set this to a smaller -#' value. +#' observations used for the approximate LOO CV (anything from 1 to the number +#' of all observations). Smaller values lead to faster computation but higher +#' uncertainty in the evaluation part. Currently, subsampled PSIS-LOO CV +#' should only be used for faster experimentation. If `NULL`, all observations +#' are used (as by default). #' @param K Only relevant if `cv_method = "kfold"` and if `cvfits` is `NULL` #' (which is the case for reference model objects created by #' [get_refmodel.stanreg()] or [brms::get_refmodel.brmsfit()]). Number of diff --git a/man/cv_varsel.Rd b/man/cv_varsel.Rd index f15ae1deb..a818f2cae 100644 --- a/man/cv_varsel.Rd +++ b/man/cv_varsel.Rd @@ -66,11 +66,11 @@ contrast to a standard LOO CV). In the \code{"kfold"} case, a \eqn{K}-fold CV is performed. See also section "Note" below.} \item{nloo}{\strong{Caution:} Still experimental. Only relevant if \code{cv_method = "LOO"}. Number of subsampled PSIS-LOO CV folds, i.e., number of -observations used for the approximate LOO CV (anything between 1 and the -original number of observations). Smaller values lead to faster computation -but higher uncertainty in the evaluation part. If \code{NULL}, all observations -are used, but for faster experimentation, one can set this to a smaller -value.} +observations used for the approximate LOO CV (anything from 1 to the number +of all observations). Smaller values lead to faster computation but higher +uncertainty in the evaluation part. Currently, subsampled PSIS-LOO CV +should only be used for faster experimentation. If \code{NULL}, all observations +are used (as by default).} \item{K}{Only relevant if \code{cv_method = "kfold"} and if \code{cvfits} is \code{NULL} (which is the case for reference model objects created by