Abstract
In this paper a general class of nonparametric test statistics, which includes both linear and nonlinear rank tests for the accelerated failure time model, is inverted into estimating equations for multiple regression. For right-censored data this general class of semiparametric regression procedures includes the linear rank estimators of Tsiatis (1990), extends the Theil-Sen estimator based on Kendall's to multiple regression, and introduces several new families of regression methods based on inverting nonlinear rank tests. These new families include the weighted generalised logrank estimators and the weighted logit-rank estimators. Several estimators of the standard errors of the regression coefficients are given. The regression coefficient estimators are consistent and asymptotically normal with variances that can be consistently estimated. Several linear and nonlinear rank-based estimators of the regression parameters and several methods of estimating their standard errors and the corresponding confidence intervals are compared in a small sample simulation in settings with and without outliers among the covariates. In these simulations the generalised logrank estimators performed well as compared to the logrank estimators when there was no outlier among the covariates and had less bias than the logrank estimators when covariate outliers existed.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.