A0399
Title: De-aliasing in two-level factorial designs: A Bayesian approach
Authors: Ming-Chung Chang - Academia Sinica (Taiwan) [presenting]
Abstract: Under limited resources on conducting follow-up trials, the inability to disentangle aliased factorial effects hinders the ubiquitous practicality of regular fractional factorial designs in the analysis of experiments. Some frequentist remedies for de-aliasing could misunderstand the underlying system behind the data. Such misinterpretation can be serious if the purpose of experimentation is to find out the mechanism in a process rather than making predictions. A Bayesian approach is proposed for de-aliasing in two-level regular factorial designs. As shown in numerical studies, our method results in a desirable model fitting and a more reliable interpretation of data than the frequentist remedies.
