eNull: Estimation of the Empirical Null

In zrmacc/LFDR: Local False Discovery Rate Estimation

Description

Estimates the parameters of the empirical null distribution using observations within a "null neighborhood". Observations in the null neighborhood are assumed to arise from the null component of the mixture distribution. The null neighborhood is specified using either a lower and upper bound, or the central proportion of z scores supposed to arise from the null.

Usage

eNull(z, L = NULL, U = NULL, p = NULL, method = "ML", f = NULL)

Arguments

z	Observations.
L	Lower limit of null neighborhood.
U	Upper limit of null neighborhood.
p	Central proportion of z scores belonging to null neighborhood.
method	Either 1. CM central matching, or 2. ML maximum likelihood
f	Estimated mixture density if method="CM".

Value

List containing the estimated mean and variance of the null component, and the proportion of observations arising from the null component.

zrmacc/LFDR documentation built on May 3, 2019, 9:01 p.m.