Bayesian Variable and Transformation Selection in Linear Regression.

Presents information on a study which proposed a Markov chain Monte Carlo model composition method which allows to average over linear regression models when the space of models under consideration is very large. Utilization of Box-Cox class of transformations for model and predictor interpretation; Discussion on the Bayesian framework and selection of prior distributions; Use of the Markov chain Monte Carlo model composition for Bayesian model averaging.