stressplot.censorReg(object, p = c(0.01, 0.05, 0.1, 0.5), method="KM", main.title=<<see below>>, sub=<<see below>>, xlab="Time to Failure", ylab=NULL, units=NULL, add=F, lty=1:length(p), col=1 + 1:length(p), xlim=NULL, ylim=NULL, add.data=T, jitter.data=F, add.legend=F, legend.loc=NULL)
"censorReg"
or of class
"censorRegList"
with exactly one
independent or stratification variable (offset variables are okay).
"x"
, and
let the response be
"censor(y)"
.
Then the possible methods for these alternative models are:
"KM" censor(y) ~ strata(x) Turnbull's generalization of
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
"none" The comparison points are not
plotted.
The default is method="KM".
paste("S/N Plot for ", deparse(substitute(object)), "\n", "method = ",
method2, "(", method, "), ", distribution2, "(", link2, ")", sep = "")
attr(object, "date")
.
TRUE
, data are displayed on the plot.
TRUE
, jittered data are displayed on the plot.
TRUE
, a legend is added to the plot
(using the
locator
function)
giving the quantiles associated with each line and color type.
NULL
and
add.legend=T
,
the
locator
function is used
so you must click in the graph at the point for locating
the upper left corner of the legend.
An option is
legend.loc="auto"
which automatically locates the legend.
A plot of the covariate versus the time to failure at the indicated quantiles is displayed.
a <- censorReg(censor(days,event)~voltage, data=capacitor2, weights=weights) stressplot.censorReg(a, method="separate") stressplot.censorReg(a)