| Type: | Package |
| Title: | Robust Likelihood Ratio Test and Confidence Intervals for the Cox Model |
| Version: | 0.1.0 |
| Author: | Yongwu Shao [aut, cre, cph] |
| Maintainer: | Yongwu Shao <ywshao@gmail.com> |
| Description: | Calculate the likelihood ratio test p-value and likelihood confidence intervals for misspecified Cox models, as described in Shao and Guo (2025) <doi:10.48550/arXiv.2508.11851>. |
| Imports: | survival |
| License: | GPL-3 |
| Encoding: | UTF-8 |
| NeedsCompilation: | no |
| Packaged: | 2026-01-16 17:40:13 UTC; oldwo |
| Repository: | CRAN |
| Date/Publication: | 2026-01-21 20:20:08 UTC |
Robust Likelihood Ratio Test and Confidence Intervals for the Cox Model
Description
Calculate the (robust) likelihood ratio test p-values and confidence intervals for the Cox model.
Usage
CoxLikelihood(time, event, X, robust = TRUE, weights = NULL, alpha = 0.05)
Arguments
time |
time of the event or censoring. |
event |
a binary variable indicating whether the record is an event or is censored. 1 is for event, 0 is for censoring. |
X |
a numeric matrix specifing the dependent variables of the Cox model. |
robust |
specifying whether the robust p-values and confidence intervals will be calculated. Default is TRUE. |
weights |
weights of each observation. The default is one for each observation. |
alpha |
1-alpha is the confidence interval (or the target coverage) of the output confidence interval. |
Details
The robust likelihood ratio test p-value is based on a scaled chi-square distribution. The robust likelihood confidence interval is generated by inverting the robust likelihood ratio test. See Shao and Guo (2026) for details.
Value
A data frame which gives the hazard ratio estimate, the robust likelihood ratio test p-values, and the robust likelihood confidence intervals.
Author(s)
Yongwu Shao
References
Shao, Yongwu, and Xu Guo. "Likelihood confidence intervals for misspecified Cox models." arXiv preprint arXiv:2508.11851 (2025).
Examples
##Create example data;
set.seed(2026);
nSubj = 100;
event = rep(1, nSubj);
X = matrix(rnorm(nSubj * 3), nSubj, 3);
time = exp(-X[,2]/2 - X[,1]^2 + X[,3]);
X = X[,-3];
## Get the robust and regular likelihood confidence intervals
CoxLikelihood(time, event, X, robust = FALSE);
CoxLikelihood(time, event, X, robust = TRUE);