lfdr {mixer}R Documentation

estimate local FDR

Description

estimate local FDR based on kernel estimation of density

Usage

lfdr(x, precision = 0.01, M=4, nL=20, nR=20)

Arguments

x a vector of signals/scores
precision because the non-parametric methos is used, local FDR is estimated at fixed points, precision is the gap between adjacent points where local FDR are estimated. Currently, precision can only be one of 0.01 (default), 0.001, or 0.0001.
M parameter used by function savgol.coef to identify the modes of density(x). M is the order of the polynomial used to approximate the desired function. Allowed values of M are 4, 5, or 6
nL parameter used by function savgol.coef to identify the modes of density(x). Given one point X, nL is the number of data points to the left of X that are used to estimate the direvative at X
nR parameter used by function savgol.coef to identify the modes of density(x). Given one point X, nR is the number of data points to the right of X that are used to estimate the direvative at X

Value

a list of class lfdr.

mean0 mean value of null distribution
var0 variance of null distribution
pi0 estimated proportion of null
precision estimated proportion of null
pt points where local FDR is estimated
f0 estimated density of null
fb estimated density mixutre
lfdr estimated local FDR
median median(x)

Author(s)

Wei Sun

See Also

plot.lfdr


[Package mixer version 1.03 Index]