check.ps.and.zs: check.ps.and.zs

In npetraco/fdrID: Computes local false discovery rate probabilities using JAGS, Stan and code from Efron, Turnbull and Narasimhan locfdr

Description

Run diagnostic tests on p-values and z-values for both null (non-match) and non-null (match) scores.

Usage

check.ps.and.zs(null.p.values = NULL, nonnull.p.values = NULL,
  printQ = FALSE, plotQ = FALSE)

Arguments

null.p.values	Optional. The p-values from the null (non-match, KM) scores.
nonnull.p.values	Optional. The p-values from the non-null (match, KNM) scores.
plotQ	Diagnostic plots?

Details

Check to see if: a. Null z-values are = +/- inf, Non-Null z-values are = -/+. If things are operating according to plan (cf. Efron 2006) the Null z-values should be between +/- 3 with ~ 99 (i.e. z_Null ~ N(0,1) approximately). Extreme Null z-values are pathological. Check the KNM scores and/or density fit to the KNM scores for weird-ness. It may be due to overly thin tails.

Value

A list with indices of problem p/z-values and a list containing statistical test results for uniformity(p-values)/normality(z-values).

Examples

XXXX

npetraco/fdrID documentation built on May 23, 2019, 9:33 p.m.