RRV {rankreg}R Documentation

Estimate the variance of a Rank estimator in censored AFT model

Description

Compute the (direct) variance-covariance estimator of the rank regression estimator in the censored AFT model.

Usage

RRV(y, d, x, beta, type="Gehan")  

Arguments

y a vector containing the censored responses in the AFT model.
d a vector of 1's and 0's. censoring indicator. 1(uncensor), 0(censored). Both y and d should be of length n.
x the design matrix, of size n by q.
beta a vector of length q. Should be the solution of the estimation equation from rankreg().
type either equal to "Gehan" (default) or "Logrank".

Details

Use local regression to estimate the slope of the estimating function, together with the variance estimator of the estimating function (from RankRegV), it returns the estimate of the variance for the rank estimator.

Value

Two q by q matrix, the estimators of the variance-covariance matrix of beta hat. In the first estimator, we get the slope matrix by (mv) regression and then inverting it before doing some more computation to form the var-cor estimator of the beta hat. In the second estimator, we begin with the inverse (mv) regression to obtain the inverse of slope matrix directly.

Author(s)

Mai Zhou.

References

Kalbfleisch, J. and Prentice, R. (2002) {em The Statistical Analysis of Failure Time Data.} 2nd Ed. Wiley, New York. (Chapter 7)


[Package rankreg version 0.2 Index]