getPosteriorParms {bfp}

Extract updated posterior parameters for the normal inverse gamma distribution from a model, given a shrinkage factor.

Description

Conditional on a fixed shrinkage factor t=g/(g+1), the posterior joint distribution of the effects and the regression variance is normal inverse gamma. With this function, you can compute the parameters of this distribution.

Usage

getPosteriorParms(x, shrinkage=x[[1]]$postExpectedShrinkage,  
                  design = getDesignMatrix(x))

Arguments

x

a valid BayesMfp-Object, only first list element will be recognized

shrinkage

shrinkage factor used in the computations (defaults to the posterior expected shrinkage factor in the model x[1])

design

(centered) design matrix for the model

Value

A list with four parameters:

aStar

the first parameter of the inverse gamma distribution

VStar

the covariance matrix part of the multivariate normal distribution

mStar

the expectation of the multivariate normal distribution

bStar

the second parameter of the inverse gamma distribution

Author(s)

Daniel Saban\'es Bov\'e

Examples

## construct a BayesMfp object
set.seed(19)

x1 <- rnorm (n=15)
x2 <- rbinom (n=15, size=20, prob=0.5) 
x3 <- rexp (n=15)

y <- rt (n=15, df=2)

test <- BayesMfp (y ~ bfp (x2, max = 4) + uc (x1 + x3), nModels = 200, method="exhaustive")

## now get the posterior parameters of the third best model
getPosteriorParms(test[3])