View source: R/confidenceEnvelope.R
confCurveFromFam | R Documentation |
Confidence bounds on the true/false positives among most significant items
confCurveFromFam(
p.values,
refFamily,
param,
K = length(p.values),
what = c("TP", "FDP")
)
p.values |
A vector containing the p-values for all m hypotheses, sorted increasingly |
refFamily |
A character value, the reference family to be used. Should be either "Simes" (or equivalenlty, "Linear"), "Beta", or "Oracle". |
param |
A numeric value or vector of parameters for the reference family. |
K |
For JER control over |
what |
A character vector, the names of the post hoc bounds to be computed, among:
Defaults to |
param
should be a numeric value unless refFamily == "Oracle"
. In the latter case, 'param“ should be a boolean vector of
length m indicating whether each null hypothesis is true or false.
A data.frame
with m
rows and 5 columns:
x: Number of most significant items selected
family: Matches input argument refFamily
param: Matches argument param
procedure: Label for the procedure, typically of the form 'refFamily(param)'
bound: Value of the post hoc bound
stat: Type of post hoc bound, as specified by argument bound
Gilles Blanchard, Pierre Neuvial and Etienne Roquain
# Generate Gaussian data and perform multiple tests
sim <- gaussianSamples(m = 502, rho = 0.5, n = 100, pi0 = 0.8, SNR = 3, prob = 0.5)
rwt <- rowWelchTests(sim$X, sim$categ, alternative = "greater")
# calculate, print, and plot confidence bound
cb <- confCurveFromFam(rwt$p.value, refFamily = "Simes", param = 0.1)
head(cb)
plotConfCurve(cb, xmax = 200)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.