Implements Bayesian model averaging for settings with many
candidate regressors relative to the available sample size, including
cases where the number of regressors exceeds the number of observations.
By restricting attention to models with at most M regressors, the package
supports reduced model space inference, thereby preserving degrees of
freedom for estimation. It provides posterior summaries, Extreme Bounds
Analysis, model selection procedures, joint inclusion measures, and
graphical tools for exploring model probabilities, model size
distributions, and coefficient distributions. When the model space is too
large to enumerate, it can be explored by Markov chain Monte Carlo model
composition instead. The methodological approach follows Doppelhofer and
Weeks (2009)