BMS toolbox for Matlab: Bayesian Model Averaging (BMA)

Do Bayesian Model Averaging (BMA) via a hidden instance of R (Windows only).

Bayesian Model Averaging for linear models under Zellner's g prior. Options include: fixed (BRIC, UIP, ...) and flexible g priors (Empirical Bayes, hyper-g), 5 kinds of model prior concepts, and model sampling via model enumeration or MCMC samplers (Metropolis-Hastings plain or reversible jump). Post-processing allows for inference according to different concepts (likelihood vs MCMC-based) and for plotting (posterior model size and coefficient densities, best models, model convergence, BMA comparison).

Needs the R D-COM interface or RAndFriends installed.

Works for Matlab 6.5 and later

For more details see:
http://bms.zeugner.eu/matlab/