post.shrinkage.minmax: Computes shrinkage of fitted estimates over regressions

In cudaBayesreg: CUDA Parallel Implementation of a Bayesian Multilevel Model for fMRI Data Analysis

Description

post.shrinkage.minmax computes the maximum and minimum fitted estimates, as a function of the mean regression coefficient estimates over all regressions.

Usage

post.shrinkage.minmax(out, X, vreg, plot=T)

Arguments

out	output of MCMC simulation
X	regression matrix used in the simulation
vreg	number of the regression coefficient
plot	{T,F} output plot (default=T)

Details

The plot helps visualizing shrinkage by analyzing the influence of the hyperparameter nu on the dispersion of the fitted maximum and minimum estimates. Different shrinkage plots may be compared for simulations with different nu values.

Value

a list containing

yrecmin	minimum values of fitted values
yrecmax	maximum values of fitted values
beta	mean of estimated coefficients over all regressions

Author(s)

A. Ferreira da Silva, Universidade Nova de Lisboa, Faculdade de Ciencias e Tecnologia,
[e-mail].

See Also

cudaMultireg.slice, read.fmrislice

Examples

## Not run: 
slicedata <- read.fmrislice(fbase="fmri", slice=3, swap=FALSE)
ymaskdata <- premask(slicedata)
fsave <- paste(tempdir(),"/simultest1",fileext = ".sav", sep="")
nu1 <- 3
out <- cudaMultireg.slice(slicedata, ymaskdata, R=2000, keep=5, nu.e=nu1,
  fsave=fsave1, zprior=FALSE, rng=1)
vreg <- 2
post.shrinkage.minmax(out, slicedata$X, vreg=vreg) 

## End(Not run)

cudaBayesreg documentation built on May 29, 2017, 6:19 p.m.