bma_wrapper {mpower}

Fits a linear model with Bayesian model selection with significance criteria: PIP and posterior probability of nonzero coefficients being on one side of zero.

Description

Fits a linear model with Bayesian model selection with significance criteria: PIP and posterior probability of nonzero coefficients being on one side of zero.

Usage

bma_wrapper(y, x, args = list())

Arguments

y

A vector of outcome

x

A matrix of predictors

args

A list of arguments see R function 'BMA::bic.glm()'.

Value

A list of vectors whose values are between 0 and 1

beta

The smaller posterior probability of the coefficients being to one side of zero: min(Pr(beta >0), Pr(beta<0)).

pip

PIP of the effect not being zero.

time

elapsed time to fit the model.

Reference

Raftery A, Hoeting J, Volinsky C, Painter I, Yeung KY (2021).BMA: Bayesian model averaging. R package version 3.18.15.