iso.ci: Calculate confidence intervals for cumulative distribution...

In IsoCI: Confidence intervals for current status data based on transformations and bootstrap.

Description

Calculate untransformed and transforemd Wald-type confidence intervals, bootstrap confidente inverals and boostrap-Wald confidnece intervals for cumulative distribution function of current stuas failure times.

Usage

iso.ci(z, d, alpha=0.05, h.opt=0.3, nboots=500, method="wald.tr", seed=1253)

Arguments

z	a vector of covariate
d	a vector of outcome
alpha	a test level for 100*(1-alpha) two-sided confidence intervals.
h.opt	an optimal bandwidth for estimating g (for "wald.tr" only)
nboots	the number of bootstrap iteration (for "bt" and "bt.wald")
method	"wald.tr" for untransformed and transformed Wald confidence intervals, "bt" for bootstrap confidence intervals, and "bt.wald" for bootstrap-Wald confidence intervals.
seed	seed value for bootstarp and bootstrap-Wald methods

Value

z	ordered covariate
yf	NPMLE estimate for cumulative distribution function
wald.lhm	left side of non-transformed Wald CI
wald.rhm	right side of non-transformed Wald CI
logit.lhm	left side of logit transformed Wald CI
logit.rhm	right side of logit transformed Wald CI
llog.lhm	left side of log(-log) transforemd Wald CI
llog.rhm	right side of log(-log) transforemd Wald CI
nbt.lhm	left side of bootstrap CI
nbt.rhm	right side of bootstrap CI
bt.wald.lhm	left side of non-transformed boostrap-Wald CI
bt.wald.rhm	right side of non-transformed boostrap-Wald CI
bt.logit.lhm	left side of logit transformed boostrap-Wald CI
bt.logit.rhm	right side of logit transformed boostrap-Wald CI
bt.llog.lhm	left side of log(-log) transforemd boostrap-Wald CI
bt.llog.rhm	right side of log(-log) transforemd boostrap-Wald CI

Author(s)

Choi, B. Y., Fine, J. P., and Brookhart, M. A.

References

Choi, B. Y., Fine, J. P., and Brookhart, M. A. (2013) Practicable confidence intervals for current status data. Statistics in Medicine 32, 1419-1428.

Ghosh, D., Banerjee, M., and Biswas, P. (2008). Inference for Constrained Estimation of Tumor Size Distributions. Biometrics 64, 1009-1017.

Groeneboom, P. and Wellner, J. A. (1992). Information Bounds and Nonparametric Maximum Likelihood Estimation. Boston: Birkhauser.

See Also

bandwidth.choose

Examples

# simulating data


n <- 50
z <- rexp(n)
pz <- pexp(z)
d <- rbinom(n,1,pz)

# finding optimal bandwidth for estimationg a density function
h.opt = bandwidth.choose(h.set=seq(0.1,2,len=15),z=z,d=d)

# Untransforemd and transformed Wald-type confidence intervals
fit.wald <- iso.ci(z=z,d=d,h.opt=h.opt$h.opt)

# Bootstrap confidence intervals
## Not run: fit.bt <- iso.ci(z=z,d=d,method="bt",nboots=100) 

# Untransforemd and transformed bootstrap-Wald-type confidence intervals
## Not run: fit.bt.wald <- iso.ci(z=z,d=d,method="bt.wald",nboots=100)

IsoCI documentation built on May 2, 2019, 8:31 a.m.