From 737498ef4ec4a95f0322baaee242c47995975cf3 Mon Sep 17 00:00:00 2001 From: john Date: Mon, 19 Aug 2024 16:25:08 +0200 Subject: [PATCH] update note for RCLL --- R/MeasureSurvRCLL.R | 2 +- man/mlr_measures_surv.rcll.Rd | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/R/MeasureSurvRCLL.R b/R/MeasureSurvRCLL.R index 613551c2..cb961acd 100644 --- a/R/MeasureSurvRCLL.R +++ b/R/MeasureSurvRCLL.R @@ -17,7 +17,7 @@ #' RCLL is proper given that censoring and survival distribution are independent, see Rindt et al. (2022). #' #' **Note**: Even though RCLL is a proper scoring rule, the calculation of \eqn{f(t)} (which in our case is discrete, i.e. it is a *probability mass function*) for time points in the test set that don't exist in the predicted survival matrix (`distr`), results in 0 values, which are substituted by `"eps"` in our implementation, therefore skewing the result towards \eqn{-log(eps)}. -#' This is the intractable likelihood problem discussed in Rindt et al. (2022), where the authors perform interpolation to get non-zero values for the \eqn{f(t)}. +#' This problem is also discussed in Rindt et al. (2022), where the authors perform interpolation to get non-zero values for the \eqn{f(t)}. #' Until this is handled in `mlr3proba` some way, we advise against using this measure for model evaluation. #' #' @section Parameter details: diff --git a/man/mlr_measures_surv.rcll.Rd b/man/mlr_measures_surv.rcll.Rd index bfec2795..1a1a4b47 100644 --- a/man/mlr_measures_surv.rcll.Rd +++ b/man/mlr_measures_surv.rcll.Rd @@ -15,7 +15,7 @@ density function and \eqn{S} the survival function. RCLL is proper given that censoring and survival distribution are independent, see Rindt et al. (2022). \strong{Note}: Even though RCLL is a proper scoring rule, the calculation of \eqn{f(t)} (which in our case is discrete, i.e. it is a \emph{probability mass function}) for time points in the test set that don't exist in the predicted survival matrix (\code{distr}), results in 0 values, which are substituted by \code{"eps"} in our implementation, therefore skewing the result towards \eqn{-log(eps)}. -This is the intractable likelihood problem discussed in Rindt et al. (2022), where the authors perform interpolation to get non-zero values for the \eqn{f(t)}. +This problem is also discussed in Rindt et al. (2022), where the authors perform interpolation to get non-zero values for the \eqn{f(t)}. Until this is handled in \code{mlr3proba} some way, we advise against using this measure for model evaluation. } \section{Dictionary}{