Bayesian subset selection methods for finding engineering design values: an application to lumber strength (Q2316974)

New subset selection procedures are presented in the paper, which are motivated by an application (i.e., involving strength of lumber) in which lumber from different species are combined into a single species group, and a subset of that group, called the controlling subset, must be selected. The empirical results indicate the advantage of the two Bayes-based procedures over the the non-Bayesian method. Simulation studies present comparison of these two Bayesian procedures to the existing method.