Premium
The factor aliased effect number pattern and its application in experimental planning
Author(s) -
Zhou Qi,
Balakrishnan Narayanaswamy,
Zhang Runchu
Publication year - 2013
Publication title -
canadian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.804
H-Index - 51
eISSN - 1708-945X
pISSN - 0319-5724
DOI - 10.1002/cjs.11190
Subject(s) - fractional factorial design , rank (graph theory) , ranking (information retrieval) , factorial experiment , factor (programming language) , design of experiments , computer science , statistics , paired comparison , mathematics , machine learning , combinatorics , programming language
Optimality criteria are usually used to choose fractional factorial designs in applications. Within an optimal design, the effects of factors assigned to different columns may be estimated with different precisions. Among factors to be investigated in an experiment, the user often has prior information on their relative importance. Thus, it is beneficial to assign most important factors to columns enabling most precise estimation. In this paper, we introduce a criterion to rank the columns of a regular design and use the criterion to GMC designs accordingly. We study the mathematical properties of the new ranking practice and provide concrete guidance on assigning factors in some GMC designs. The Canadian Journal of Statistics 41: 540–555; 2013 © 2013 Statistical Society of Canada