unleash: UNimodally Leveraging Empirical-null with Adaptive SHrinkage

In dcgerard/unleash: UNimodally Leveraging Empirical-null with Adaptive SHrinkage

Description

This is a wrapper for ash with a variance inflation term that expands the variances to as large a value as possible while allowing the prior to be unimodal.

Usage

unleash(betahat, sebetahat, mixcompdist = c("uniform", "halfuniform",
  "normal", "+uniform", "-uniform"), df = NULL, pen = 1, maxiter = 500,
  tol = 10^-6, plot_update = FALSE, ...)

Arguments

betahat	a p vector of estimates
sebetahat	a p vector of corresponding standard errors
mixcompdist	distribution of components in mixture ("uniform","halfuniform" or "normal"; "+uniform" or "-uniform"), the default is "uniform". If you believe your effects may be asymmetric, use "halfuniform". If you want to allow only positive/negative effects use "+uniform"/"-uniform". The use of "normal" is permitted only if df=NULL.
df	appropriate degrees of freedom for (t) distribution of betahat/sebetahat, default is NULL which is actually treated as infinity (Gaussian)
pen	The variance inflation penalty.
maxiter	The maximum number of iterations to run.
tol	The stopping criterion.
plot_update	A logical. Should we plot the updates?
...	Further arguments to be passed to ash.workhorse.

Author(s)

David Gerard

dcgerard/unleash documentation built on May 23, 2019, 8:37 a.m.