Sequentially designed experiments for screening out bad models with F tests

Abstract

SUMMARY The sequential design of experiments to screen out invalid models among a set of contemplated linear models is proposed. The method is illustrated with a simple polynomial example.