surv_brier.coxex: Calculate Brier scores for survival models.

In PiotrTymoszuk/coxExtensions: Extended Inference and Goodness of Fit for Cox Models

Calculate Brier scores for survival models.

Description

The Brier score for survival models is computed as a squared distance between the predicted survival probability and and the actual 0/1-coded survival for each unique time point of the dataset.

Usage

surv_brier(fit, ...)

## S3 method for class 'coxex'
surv_brier(fit, splitMethod = "none", ...)

## S3 method for class 'coxph'
surv_brier(fit, data, splitMethod = "none", ...)

Arguments

fit	a survival model.
...	extra arguments passed to pec.
splitMethod	validation method used for calculation of the Bier scores as specified by the respective pec argument. One of 'none' (no validation), 'cvK' (K-fold cross-validation, e.g. 'cv10'), 'boot' (bootstrap), 'BootCv' (bootstrap cross-validation), 'Boot632', 'Boot632plus', 'loocv' or 'NoInf', see the upstream function for details.
data	the data frame used for the model construction.

Details

See, pec for details. The surv_brier is a S3 generic function.

Value

an object of the brier class with the plot method and multiple methods shared with traditional data frames.

References

• Graf, E., Schmoor, C., Sauerbrei, W. & Schumacher, M. Assessment and comparison of prognostic classification schemes for survival data. Stat. Med. 18, 2529–2545 (1999).

PiotrTymoszuk/coxExtensions documentation built on Feb. 6, 2024, 10:58 p.m.