RankRegV {rankreg} | R Documentation |
Compute the score type test, with the chi square value for testing Ho: betanull.
It also resturns the variance-covariance matrix of the rank estimating function (score function) at beta for the censored AFT model.
RankRegV(y, d, x, beta, betanull = beta, type="Gehan")
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. Should not have a column of 1s. |
beta |
a vector of dimension q. Usually it should be the solution
of the estimation equation, from rankreg() . |
betanull |
the null value (under H0) of beta to be tested. |
type |
either equal to "Gehan" (default) or "Logrank". |
It returns the chi square value and P-value which are based on
EF(betanull)*Vtest(beta)^(-1)* EF(betanull)
.
The input beta
usually should be the solution
of the estimating function.
This solution can be obtained by using rankaft
.
But other values of beta may also make sense sometime.
When betanull
equals beta
from rankaft
,
then the chi square
statistic should be zero (since EF(betanull)=0
)
and the P-value should be 1.
A list with VEF
, which is the variance estimator of the
estimating function at beta
;
chisqvalue
which
is the quadratic form EF(betanull)*VEF^(-1)*EF(betanull)
;
and Pval
that is the P-value from the chi square quantile.
Mai Zhou.
Kalbfleisch, J. and Prentice, R. (2002) {em The Statistical Analysis of Failure Time Data.} 2nd Ed. Wiley, New York. (Chapter 7)
data(myeloma) RankRegV(y=myeloma[,1],d=myeloma[,2],x=cbind(myeloma[,3],myeloma[,4]), beta=c(-15.01117, 1.317596), betanull=c(-15.01117, 1.317596)) ### you should get a chisquare = 0.001124574 , Pval = 0.9994379 GP1 <- c(143,164,188,188,190,192,206,209,213,216,220,227,230,234,246,265,304, 216,244) GP2 <- c(142,156,163,198,205,232,232,233,233,233,233,239,240,261,280,280,296, 296,323,204,344) status1 <- c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0,0) status2 <- c(1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0,0) RankRegV(y=c(GP1,GP2),d=c(status1,status2),x=c(rep(0,19),rep(1,21)), beta=0,betanull=0)