Software

Bayesian Model Averaging (BMA)
An Object-Oriented Package for Ox

BMA v1.0 (Bayesian Model Averaging) is a package, written in Ox, for conducting inference over a potentially large space of linear regression models. Operationalising the MC3 algorithm of Madigan and York (1995), BMA represents an important development for both the OxMetrics and, more generally, model averaging computing environment by making available a comprehensive model averaging toolset with a user friendly interface.

Users may access the functions within BMA by writing a short Ox programs which creates an object of the BMA class. We demonstrate the capabilities of BMA by using a number of well known applications from the literature. This document will serve as a manual for BMA

Gernot Doppelhofer
Norwegian School of Economics and Business Administration (NHH)

Dr. Melvyn Weeks
Faculty of Economics, University of Cambridge