probplot6.censorReg(object, plot.at=NULL, method=c("KM", "separate", "factor",
"regression", "one", "none"), distribution=c("weibull","lognormal",
"loglogistic"), mle.interval=F, conf.level=0.95, xlab=NULL,
ylab="Probability", main.title=NULL, xlim=NULL, ylim=NULL,
add.legend=F)
"censorReg".
object.
x, and let the response be
censor(y). Then the possible methods for these alternative models are:
"KM" censor(y) ~ strata(x) Kaplan-Meier estimates
"separate" censor(y) ~ strata(x) censorReg estimates with
separate location and scale
for each covariate value
"factor" censor(y) ~ factor(x) censorReg estimates with
separate location and the same
scale for each covariate value
"regression" censor(y) ~ x censorReg model with location
predicted by regression on the
covariate x
"one" censor(y) ~ 1 censorReg model with no covariates
"none" no points are plotted
The default is method="KM".
response(object).
A probability plot is produced on the current graphics device. This is a multiple figure plot with up to six figures, one for each distribution.
probplot6.censorReg(censorReg(censor(days,event)~voltage, data=capacitor2, weights =
weights))