Skip to contents

Calculates the right-censored logarithmic (log), loss.

The RCLL, in the context of probabilistic predictions, is defined by $$L(f, t, \Delta) = -log(\Delta f(t) + (1 - \Delta) S(t))$$ where \(\Delta\) is the censoring indicator.

Parameters

  • eps (numeric(1)) - Value to set zero-valued scores to prevent log(0) errors, default 1e-15.

  • se (logical(1)) - If TRUE then returns standard error of the loss otherwise returns mean across all individual scores.

  • ERV (logical(1)) - If TRUE then the Explained Residual Variation method is applied, which means the score is standardised against a Kaplan-Meier baseline.

  • na.rm (logical(1)) - If TRUE (default) then removes any NAs in individual score calculations.

Dictionary

This Measure can be instantiated via the dictionary mlr_measures or with the associated sugar function msr():

MeasureSurvRCLL$new()
mlr_measures$get("surv.rcll")
msr("surv.rcll")

Meta Information

  • Type: "surv"

  • Range: \([0, \infty)\)

  • Minimize: TRUE

  • Required prediction: distr

References

Avati, A., Duan, T., Zhou, S., Jung, K., Shah, N. H., & Ng, A. (2018). Countdown Regression: Sharp and Calibrated Survival Predictions. http://arxiv.org/abs/1806.08324

Super classes

mlr3::Measure -> mlr3proba::MeasureSurv -> MeasureSurvRCLL

Methods

Inherited methods


Method new()

Creates a new instance of this R6 class.

Usage

MeasureSurvRCLL$new(ERV = FALSE)

Arguments

ERV

(logical(1))
Standardize measure against a Kaplan-Meier baseline (Explained Residual Variation)


Method clone()

The objects of this class are cloneable with this method.

Usage

MeasureSurvRCLL$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.