文摘
Suppose a large number of two-level factors is examined in an experimental situation. Under the assumption of effect sparsity it is often anticipated that only a few experimental factors play an important role in the experiment. Usually, it is not known which columns of the experimental design will be of further interest. After the initial stage of factor screening, experimenters are usually interested in identifying and estimating interactions of factors that have been found active. It is therefore practical to select a design suitable to provide essential information for this purpose, for whatever choice of active factors.