Approximating Maximally Selected Statistics

Approximates the probability that a maximally selected rank statistic is greater or equal to b.

pmaxstat(b, scores, msample, quant=FALSE)
qmaxstat(p, scores, msample)

Arguments

b

quantile.

p

propability.

scores

integer valued scores assigned to the observations.

msample

all possible splitpoints.

quant

logical. Returns the results of SR instead of P-values. Only to be used in qmaxstat.

Details

Small sample approximation by Hothorn and Lausen (2003).

Value

An upper limit for the probability that, under the hypothesis of independence, a maximally selected statistic greater equal b is observed. qmaxstat needs optimization.

References

Hothorn, T. and Lausen, B. (2003). On the Exact Distribution of Maximally Selected Rank Statistics. Computational Statistics & Data Analysis, 43, 121–137.

Examples

pmaxstat(2.5, 1:20, 5:15)
#> [1] 0.09403538