c_index: Compute the Harrell's Concordance index

In survex: Explainable Machine Learning in Survival Analysis

Compute the Harrell's Concordance index

Description

A function to compute the Harrell's concordance index of a survival model.

Usage

c_index(y_true = NULL, risk = NULL, surv = NULL, times = NULL)

Arguments

y_true	a survival::Surv object containing the times and statuses of observations for which the metric will be evaluated
risk	a numeric vector of risk scores corresponding to each observation
surv	ignored, left for compatibility with other metrics
times	ignored, left for compatibility with other metrics

Value

numeric from 0 to 1, higher values indicate better performance

References

• [1] Harrell, F.E., Jr., et al. "Regression modelling strategies for improved prognostic prediction." Statistics in Medicine 3.2 (1984): 143-152.

See Also

loss_one_minus_c_index()

Examples

library(survival)
library(survex)

rotterdam <- survival::rotterdam
rotterdam$year <- NULL
cox_rotterdam_rec <- coxph(Surv(rtime, recur) ~ .,
    data = rotterdam,
    model = TRUE, x = TRUE, y = TRUE
)
coxph_explainer <- explain(cox_rotterdam_rec)

risk <- coxph_explainer$predict_function(coxph_explainer$model, coxph_explainer$data)
c_index(y_true = coxph_explainer$y, risk = risk)

survex documentation built on Oct. 25, 2023, 1:06 a.m.