Permutation Test

Perform permutation test for a linear statistic

doTest(object, teststat = c("maximum", "quadratic", "scalar"),
       alternative = c("two.sided", "less", "greater"), pvalue = TRUE,
       lower = FALSE, log = FALSE, PermutedStatistics = FALSE,
       minbucket = 10L, ordered = TRUE, maxselect = object$Xfactor,
       pargs = GenzBretz())

Arguments

object

an object returned by LinStatExpCov.

teststat

type of test statistic to use.

alternative

alternative for scalar or maximum-type statistics.

pvalue

a logical indicating if a p-value shall be computed.

lower

a logical indicating if a p-value (lower is FALSE) or 1 - p-value (lower is TRUE) shall be returned.

log

a logical, if TRUE probabilities are log-probabilities.

PermutedStatistics

a logical, return permuted test statistics.

minbucket

minimum weight in either of two groups for maximally selected statistics.

ordered

a logical, if TRUE maximally selected statistics assume that the cutpoints are ordered.

maxselect

a logical, if TRUE maximally selected statistics are computed. This requires that X was an implicitly defined design matrix in LinStatExpCov.

pargs

arguments as in GenzBretz.

Details

Computes a test statistic, a corresponding p-value and, optionally, cutpoints for maximally selected statistics.

Value

A list.