Regression approach to estimate OF of the E(S2) values for supersaturated designs

  • CO Todo
  • JI Mbegbu


The E(s2) criterion measures the average correlation among the columns of the design matrix for supersaturated designs (SSDs). In this light, Bulutoglu and Cheng (2004) constructed some SSD’s, and then obtained their corresponding values of E(s2) by explicit computational method. In this paper however, we estimated the values of E(s2) by a regression method using existing SSDs. The results obtained were in line with the results of Bulutoglu and Cheng (2004) and the lower bound derived by Nguyen (1996) and Tang and Wu (1997) were satisfied.

Keywords: E(s2) criterion, supersaturated designs, regression, lower bound

International Journal of Natural and Applied Sciences, 7(1): 54 - 57, 2011

Author Biographies

CO Todo
Department of Statistics, School of Computing and IT, Delta State Polytechnic, Otefe-Oghara, Nigeria
JI Mbegbu
Department of Mathematics, University of Benin, Benin City, Nigeria

Journal Identifiers

eISSN: 0794-4713