Obtain posterior model probabilities after running Bayesian model selection
Usage
postProb(object, nmax, method='norm')
Arguments
object
Object of class msfit returned by modelSelection
or class mixturebf, e.g. returned by bfnormmix
nmax
Maximum number of models to report (defaults to no max)
method
Only when class(object) is msfit.
For 'norm' probabilities are obtained by renormalizing the
stored integrated likelihoods, for 'exact' they are given by the proportion
of MCMC visits to each model. 'norm' has less variability but can be biased
if the chain has not converged.
Value
A data.frame with posterior model probabilities in column pp.
Column modelid indicates the indexes of the selected covariates (empty
for the null model with no covariates).