| plot.proprate2.fit {surv2sample} | R Documentation |
This function plots estimates of cumulative rates (cumulative hazards or odds functions) for two samples of censored data. It may plot both separate estimates from the two samples and estimates based on the proportional rate model.
## S3 method for class 'proprate2.fit':
plot(x, log.transform = FALSE, diff = FALSE, lwds = 1, cols = 1,
ltys, ...)
x |
a "proprate2.fit" object, as returned by proprate2. |
log.transform |
logical. Should the logarithms of cumulative rates be plotted? |
diff |
logical. Instead of two curves, should the difference of their logarithms be plotted? |
lwds, cols, ltys |
vectors of length equal to the number of curves in plots
(4 if diff is FALSE, 2 if TRUE). These give line widths,
colours and line types for each curve. If of length 1, the value is replicated. |
... |
further plotting parameters. |
If diff is FALSE, four curves are plotted (two individual sample
estimates and two model based estimates). In this case, ltys defaults
to c(1,1,2,2). If diff is FALSE,
the function plots their differences. Then ltys defaults to c(1,1).
To omit a curve, set the corresponding component of lty to 0.
Using these plots one may visually assess the validity of the proportional rate assumption.
David Kraus (http://www.davidkraus.net/)
proprate2 for estimation
proprate2.neyman, proprate2.ks,
proprate2.gs for tests of the proportional rate
assumption
## chronic active hepatitis data
data(hepatitis)
## fit the proportional odds model
fit = with(hepatitis, proprate2(Surv(time, status), treatment,
model = 1))
## plot model-based and model-free estimates of odds functions
plot(fit)
## their logarithms
plot(fit, log.transform = TRUE)
## differences of log-functions
plot(fit, diff = TRUE)