Improving the Accuracy of Parameter Estimation of Proportional Hazards Regression with Kernel Resampling

Abstract

The accuracy of parameter estimation of proportional hazards regression (PHR) has been a concern. To improve the accuracy of the estimation, the bootstrap has been used; unfortunately, prior research revealed inconsistent findings. The current study applies a new resampling method, the kernel resampling technique (KRT), to PHR. Two empirical datasets were employed to cross-validate and compare the accuracy and stability of the estimation results through multiple replications from KRT with those from the naïve bootstrap as well as the maximum likelihood method. The study results revealed that KRT outperformed the bootstrap and maximum likelihood method in estimating parameters of PHR. The application of KRT to PHR improved the accuracy of the parameter estimation.