BaySIC (version 1.0)

fuzzy.FDR.approx: Generate Approximate Fuzzy Rejection Probabilites

Description

For hypothesis tests with discrete reference distributions, obtains fuzzy rejection probabilites for a given level of false discovery rate control

Usage

fuzzy.FDR.approx(pprev, p, alpha, N)

Arguments

pprev

numeric vector of p-values of length $l$, corresponding to strict inequality of test statistic values in a one-sided test (i.e., $P(T>t)$)

p

length $l$ numeric vector of p-values corresponding to traditional one-sided test (i.e., $P(T\geq t)$).

alpha

FDR level of interest (under Benjamini-Hochberg FDR procedure)

N

Number of Monte Carlo samples used to generate fuzzy rejection probability approximations.

Value

$l$ corresponding to the fuzzy rejection probabilities of the hypotheses represented in pprev and p, under FDR level alpha

Details

This is a Monte Carlo implementation of the fuzzy FDR work developed by Kulinskaya et al. (2007)

References

http://www.bgx.org.uk/alex/Kulinskaya-Lewin-resubmit.pdf