maxstat.test {maxstat}R Documentation

Maximally Selected Rank and Gauss Statistics

Description

Performs a maximally selected rank or Gauss test on bivariate observations.

Usage

maxstat.test(x, y, cens = NULL, smethod=c("Gauss", "Wilcoxon", "Median",
            "NormalQuantil","LogRank"), pmethod=c("none", "Lau92",
            "Lau94", "exactGauss", "HL", "min"),
            minprop = 0.1, maxprop=0.9, plot=F, xlab=NULL,...)

Arguments

x vector of data values, independent variable.
y numeric vector of data values, dependent variable.
cens censoring: 0 cencored, 1 dead (only meaningful with LogRank)
smethod kind of statistic to be computed.
pmethod kind of p-value approximation to be used.
minprop at least minprop*100% of the observations in the first group.
maxprop not more than minprop*100% of the observations in the first group.
plot logical. Should the process be plotted?
xlab use alternative xlab
... additional plotting parameters

Details

The assessment of the predictive power of a variable x for a dependent variable y can be determined by a maximally selected rank or Gauss test.

smethod determines the kind of statistic to be used. Gauss means a maximally selected Gauss statistic, Wilcoxon and Median simply denote Wilcoxon and Median scores. NormalQuantile and LogRank denote v.d. Waerden and Log-rank scores.

pmethod specifies which kind of approximation of the p-value should be used. Lau92 means the limiting distribution by a Brownian bridge (see pLausen92), Lau94 the approximation based on an improved Bonferroni inequality (see pLausen94).

exactGauss returns the exact p-value for a maximally selected Gauss statistic.

HL is a small sample approximation based on the Streitberg-R"ohmel algorithm (see pperm) introduced by Hothorn & Lausen (2001). This requires integer valued scores. For v. d. Waerden and Log-rank scores we try to find integer valued scores having the same shape. This results in slightly different scores (and a different test), the procedure is described in Hothorn (2001).

All the approximations are known to be conservative, so min gives the minimum p-value of all procedures.

Value

A list of class htest containing the following components:

statistic the value of the test statistic.
p.value the p-value for the test.
method the type of test and p-value approximation applied.
estimate the estimated cutpoint (of x) which separates y best.

Author(s)

Torsten Hothorn <Torsten.Hothorn@rzmail.uni-erlangen.de>

References

Hothorn, T. and Lausen, B. (2001). On the Exact Distribution of Maximally Selected Rank Statistics. submitted

Lausen, B. and Schumacher, M. (1992). Maximally Selected Rank Statistics. Biometrics, 48, 73–85

Hothorn, T. (2001). On Exact Rank Tests in R. R News, 1, 11–12

Examples


x <- sort(runif(20))
y <- c(rnorm(10), rnorm(10, 2))

maxstat.test(x, y, smethod="Wilcoxon", pmethod="HL", minprop=0.25, maxprop=0.75)