Average c-Efficiency

by Charles F. Dunkl and Donald E. Ramirez .

Abstract: This paper introduces, for an experimental design with design matrix X, the optimal design criterion ACE based on the integrated average of the standard errors for tests H:u'b=0, where the distribution of u satisfies a normal distribution with mean 0 and with a user specified dispersion matrix. When the dispersion matrix is the identity matrix, the criterion is equivalent to the integrated average of the standard errors with u uniformly distributed over the sphere.

Key Words: optimal design, ACE, expected radius, points of influence, integrated average, ellipsoids.

Charles F. Dunkl, cfd5z@virginia.edu
Donald E. Ramirez, der@virginia.edu

Editor: Johan O. Fellman , fellman@shh.fi

READING THE ARTICLE: You can read the article in portable document (.pdf) format (433467 bytes.)

NOTE: The content of this article is the intellectual property of the authors, who retains all rights to future publication.

This page has been accessed 2896 times since July 24, 2006.

Return to the InterStat Home Page.