MAMSE (version 0.1-3)

WKME: Kaplan-Meier Estimate

Description

Computes the weighted Kaplan-Meier estimate over some time points with optional confidence intervals.

Usage

WKME(x,ub,lb=0,time=NULL,boot=NULL,REP=1000)

Arguments

x

A list of m samples. Each element is an n by 2 matrix whose second column is an indicator of whether the time in column 1 is observed (1) or censored (0).

lb,ub

Lower and upper bounds of the integral of the MAMSE criterion.

time

A vector of times at which to compute the Kaplan-Meier estimate.

boot

When NULL, bootstrap confidence intervals are not generated. Otherwise must be a number in (0,1) corresponding to the coverage probability of the bootstrap intervals to be built.

REP

When bootstrap is used, controls the number of pseudo-sample to generate.

Value

x

Sorted list of the times (observed and censored) from each samples

weight

The size of the jump that the Kaplan-Meier estimate allocates to each time in x.

time

Vector of time points where the function is evaluated.

kme

The Kaplan-Meier estimate for Population 1 evaluated at time.

kmeCI

Pointwise bootstrap confidence interval for kme.

wkme

The weighted Kaplan-Meier estimate evaluated at time.

wkmeCI

Pointwise bootstrap confidence interval for wkme.

Details

This function calculates the weighted Kaplan-Meier estimate and can provide pointwise bootstrap confidence intervals.

References

J.-F. Plante (2007). Adaptive Likelihood Weights and Mixtures of Empirical Distributions. Unpublished doctoral dissertation, University of British Columbia.

J.-F. Plante (2009). About an adaptively weighted Kaplan-Meier estimate. Lifetime Data Analysis, 15, 295-315.

See Also

MAMSE-package, WKME.

Examples

set.seed(2009)
x=list(
  cbind(rexp(20),sample(c(0,1),20,replace=TRUE)),
  cbind(rexp(50),sample(c(0,1),50,replace=TRUE)),
  cbind(rexp(100),sample(c(0,1),100,replace=TRUE))
)

allx=pmin(1,c(x[[1]][x[[1]][,2]==1,1],x[[2]][x[[2]][,2]==1,1],
   x[[3]][x[[3]][,2]==1,1]))
K=WKME(x,1,time=sort(unique(c(0,1,allx,allx-.0001))),boot=.9,REP=100)
# Only 100 bootstrap repetitions were used to get a fast enough
# calculation on a CRAN check.

plot(K$time,K$wkme,type='l',col="blue",xlab="x",
   ylab="P(X<=x)",ylim=c(0,1))
lines(K$time,K$kme[,1],col="red")

lines(K$time,K$wkmeCI[1,],lty=2,col="blue")
lines(K$time,K$wkmeCI[2,],lty=2,col="blue")

lines(K$time,K$kmeCI[1,],lty=2,col="red")
lines(K$time,K$kmeCI[2,],lty=2,col="red")
legend(.1,.9,c("Weighted Kaplan-Meier","Kaplan-Meier"),
   col=c("blue","red"),lty=c(1,1))