Calculates the cross-entropy, or logarithmic (log), loss.

The logloss, in the context of probabilistic predictions, is defined as the negative log probability density function, \(f\), evaluated at the observation time, \(t\), $$L(f, t) = -log(f(t))$$

Censored observations in the test set are ignored.

Format

R6::R6Class() inheriting from MeasureSurv.

Construction

MeasureSurvLogloss$new(eps = 1e-15)
mlr_measures$get("surv.logloss")
msr("surv.logloss")
  • eps :: numeric(1)
    Very small number to set zero-valued predicted probabilities to, in order to prevent errors in log(0) calculation.

Meta Information

  • Type: "surv"

  • Range: \([0, \infty)\)

  • Minimize: TRUE

  • Required prediction: distr

Fields

See MeasureSurv, as well as all variables passed to the constructor.

As well as

  • eps :: numeric(1)
    Very small number to set zero-valued predicted probabilities to, in order to prevent errors in log(0) calculation.

See also