discotime.metrics.brier_score module

class discotime.metrics.brier_score.BrierScore(timepoints: _SupportsArray[dtype[Any]] | _NestedSequence[_SupportsArray[dtype[Any]]] | bool | int | float | complex | str | bytes | _NestedSequence[bool | int | float | complex | str | bytes], causes: list[int] | None = None)[source]

Bases: object

Brier score of survival model with right-censored data.

In the case of right-censored data, possibly unknown time-dependent event status can be replaced with jackknife pseudovalues from the marginal Aalen-Johansen estimates [1].

Parameters:

  • timepoints – sequence of t timepoints to evaluate.

  • causes – causes for which the brier score is calculated.

Refs:

[1] Cortese, Giuliana, Thomas A. Gerds, and Per K. Andersen. “Comparing predictions among competing risks models with time‐dependent covariates.” Statistics in medicine 32.18 (2013): 3089-3101.

class Results(null, model)

Bases: tuple

model

Alias for field number 1

null

Alias for field number 0