valorate.plot.subpop: PLOT ALL ESTIMATED LOG-RANK DISTRIBUTIONS

In valorate: Velocity and Accuracy of the LOg-RAnk TEst

Description

Plots all log-rank distributions estimated with the same object (different values of n1). This family of plots is commonly used to compare the estimated distributions.

Usage

valorate.plot.subpop.empirical(vro, which, 
  type, log, xlim, smooth, legends, 
  density, ylim, ...)

valorate.plot.subpop.empirical.to.0(vro, which, 
  type, log, xlim, smooth, legends, density, ylim, ...)

valorate.plot.subpop.empirical.scaled(vro, which, 
  type, log, xlim, 
  smooth, legends, density, ylim, 
  scale.point, ...)

Arguments

vro	the valorate object.
which	The values of n1 that will be shown. NULL to plot them all.
type	typical plot parameter: "p"=points, "l"=lines (default), "o"=overlap.
log	typical plot parameter : specify which axis are shown in logarithm base 10.
xlim	typical plot parameter.
ylim	typical plot parameter.
smooth	the strength of density smoothing for display purposes. The default is 10.
legends	the number of columns in legends. 0 to omit legends.
density	indicates whether all curves should represent density (default to TRUE). FALSE to scale to maximum.
...	arguments passed to plot.
scale.point	a double between 0 and 0.5 (exclusive) that determines the two points in quantiles in which all densities will be 'equalized'. The quantiles are scale.point and 1-scale.point. 0.5 should be avoided.

Details

valorate.plot.subpop.empirical plots all log-rank distributions estimated with the same object (different values of n1) in raw densities and scales. valorate.plot.subpop.empirical.to.0 is similar to valorate.plot.subpop.empirical but shift distributions to 0 and scale horizontal axis to similar limits. valorate.plot.subpop.empirical.scaled is similar to valorate.plot.subpop.empirical but scales the distributions to have the same scale.point(s) (in x) for all distributions. It also shifts all distribution to zero. This helps to compare the tendencies of the overall distributions.

Value

Nothing.

Author(s)

Victor Trevino [e-mail]

References

Trevino et al. 2016 http://bioinformatica.mty.itesm.mx/valorateR

See Also

new.valorate. valorate.survdiff.

Examples

## Create a random population of 100 subjects 
## having 20 events
subjects <- numeric(100)
subjects[sample(100,20)] <- 1
vo <- new.valorate(rank=subjects, sampling.size=100000, verbose=TRUE)

for (i in c(5,10,20,30,40,50)) {
  groups <- numeric(100)
  groups[sample(100,i)] <- 1 
  valorate.survdiff(vo, groups) 
}

## Not run: valorate.plot.subpop.empirical(vo)
## Not run: valorate.plot.subpop.empirical.to.0(vo)

valorate documentation built on May 1, 2019, 9:10 p.m.