hc.score: Compute Empirical Higher Criticism Scores and Corresponding...

In fdrtool: Estimation of (Local) False Discovery Rates and Higher Criticism

Compute Empirical Higher Criticism Scores and Corresponding Decision Threshold From p-Values

Description

hc.score computes the empirical higher criticism (HC) scores from p-values.

hc.thresh determines the HC decision threshold by searching for the p-value with the maximum HC score.

Usage

hc.score(pval)
hc.thresh(pval, alpha0=1, plot=TRUE)

Arguments

pval	vector of p-values.
alpha0	look only at a fraction alpha0 of the p-values (default: 1, i.e. all p-values).
plot	show plot with HC decision threshold.

Details

Higher Criticism (HC) provides an alternative means to determine decision thresholds for signal identification, especially if the signal is rare and weak.

See Donoho and Jin (2008) for details of this approach and Klaus and Strimmer (2012) for a review and connections with FDR methdology.

Value

hc.score returns a vector with the HC score corresponding to each p-value.

hc.thresh returns the p-value corresponding to the maximum HC score.

Author(s)

Bernd Klaus and Korbinian Strimmer (https://strimmerlab.github.io).

References

Donoho, D. and J. Jin. (2008). Higher criticism thresholding: optimal feature selection when useful features are rare and weak. Proc. Natl. Acad. Sci. USA 105:14790-15795.

Klaus, B., and K. Strimmer (2013). Signal identification for rare and weak features: higher criticism or false discovery rates? Biostatistics 14: 129-143. <DOI:10.1093/biostatistics/kxs030>

See Also

fdrtool.

Examples

# load "fdrtool" library
library("fdrtool")

# some p-values
data(pvalues)

# compute HC scores
hc.score(pvalues)

# determine HC threshold
hc.thresh(pvalues)

fdrtool documentation built on Sept. 11, 2024, 9:10 p.m.