Coherent dispersion criteria for optimal experimental design. (English) Zbl 0948.62057

Dawid, A. Philip; Sebastiani, Paola

Ann. Stat. 27, No. 1, 65-81 (1999).

Authors’ abstract: We characterize those coherent design criteria which depend only on the dispersion matrix (assumed proper and nonsingular) of the “state of nature”, which may be a parameter-vector or a set of future observables and describe the associated decision problems. Connections are established with the classical approach to optimal design theory for the normal linear model, based on concave functions of the information matrix. Implications of the theory for more general models are also considered.
Comments: This is a paper for theoretical workers in optimal design and analysis of experiments. It reaches aspects of both the classical Kiefer’s optimal design methods and Bayesian approaches.

Reviewer: B.L.Raktoe (Singapore)

MSC:

62K05	Optimal statistical designs
62C10	Bayesian problems; characterization of Bayes procedures

Keywords:

dispersion standard; Kiefer’s optimal design

References:

[1]	Aitchison, J. and Dunsmore, I. R. (1975). Statistical Prediction Analysis. Cambridge Univ. Press. · Zbl 0327.62043
[2]	Atkinson, A. C. (1996). The usefulness of optimum experimental designs (with discussion). J. Roy. Statist. Soc. Ser. C 58 59-76; 95-111. JSTOR: · Zbl 0850.62602
[3]	Atkinson, A. C. and Donev, A. N. (1992). Optimum Experimental Design. Clarendon Press, Oxford. · Zbl 0829.62070
[4]	Bates, R. A., Buck, R. J., Riccomagno, E. and Wynn, H. P. (1996). Experimental design and observation for large systems (with discussion). J. Roy. Statist. Soc. Ser. B 58 77-111. JSTOR: · Zbl 0850.62627
[5]	Bernardo, J. M. and Smith, A. F. M. (1994). Bayesian Theory. Wiley, New York. · Zbl 0796.62002
[6]	Chaloner, K. (1984). Optimal Bayesian experimental designs for linear models. Ann. Statist. 12, 283-300. · Zbl 0554.62063 · doi:10.1214/aos/1176346407
[7]	Chaloner, K. and Verdinelli, I. (1995). Bayesian experimental design: a review. Statist. Sci. 10 273-304. · Zbl 0955.62617 · doi:10.1214/ss/1177009939
[8]	Dawid, A. P. (1994). Proper measures of discrepancy, uncertainty and dependence, with applications to predictive experimental design. Research Report 139, Dept. Statistical Science, Univ. College London. · doi:10.1214/lnms/1215463797
[9]	DeGroot, M. H. (1970). Optimal Statistical Decisions. McGraw-Hill, New York. · Zbl 0225.62006
[10]	Eaton, M. L., Giovagnoli, A. and Sebastiani, P. (1994). A predictive approach to the Bayesian design problem with application to normal regression models. Technical Report 598, School of Statistics, Univ. Minnesota. · Zbl 0865.62001
[11]	Eaton, M. L., Giovagnoli, A. and Sebastiani, P. (1996). A predictive approach to the Bayesian design problem with application to normal regression models. Biometrika 83 111-125. JSTOR: · Zbl 0865.62001 · doi:10.1093/biomet/83.1.111
[12]	Firth, D. (1996). Discussion of the papers by Atkinson (1996) and by Bates et al. (1996). J. Roy. Statist. Soc. Ser. B 58 100. Firth, D. and Hinde, J. P. (1997a). On Bayesian D-optimum design criteria and the equivalence theorem in nonlinear models. J. Roy. Statist. Soc. Ser. B 59 793-797. Firth, D. and Hinde, J. P. (1997b). Parameter-neutral optimum design for non-linear models. J. Roy. Statist. Soc. Ser. B 59 799-811. JSTOR: · Zbl 0886.62074
[13]	Good, I. J. (1952). Rational decisions. J. Roy. Statist. Soc. Ser. B 14 107-114. JSTOR:
[14]	Kiefer, J. C. (1959). Optimum experimental designs (with discussion). J. Roy. Statist. Soc. Ser. B 21 272-319. JSTOR: · Zbl 0108.15303
[15]	Kiefer, J. C. (1974). General equivalence theorem for optimum designs: approximate theory. Ann. Statist. 2 849-879. · Zbl 0291.62093 · doi:10.1214/aos/1176342810
[16]	Lindley, D. V. (1956). On a measure of information provided by an experiment. Ann. Math. Statist. 27 986-1005. · Zbl 0073.14103 · doi:10.1214/aoms/1177728069
[17]	Lindley, D. V. (1972). Bayesian Statistics A Review. SIAM, Philadelphia. · Zbl 0246.62009
[18]	Pilz, J. (1991). Bayesian Estimation and Experimental Design in Linear Regression Models. Wiley, New York. · Zbl 0745.62068
[19]	Pronzato, L. and Walter, E. (1985). Robust experimental design via stochastic approximations. Math. Biosci. 75 103-120. · Zbl 0593.62070 · doi:10.1016/0025-5564(85)90068-9
[20]	Pukelsheim, F. (1993). Optimal Design of Experiments. Wiley, New York. · Zbl 0834.62068
[21]	Raiffa, H. A. and Schlaifer, R. S. (1961). Applied Statistical Decision Theory. MIT Press. · Zbl 0952.62008
[22]	Rockafellar, R. T. (1970). Convex Analysis. Princeton Univ. Press. · Zbl 0193.18401
[23]	Sacks, J., Welch, W., Mitchell, T. and Wynn, H. P. (1989). Design and analysis of computer experiments (with discussion). Statist. Sci. 4 409-423. · Zbl 0955.62619 · doi:10.1214/ss/1177012413
[24]	Savage, L. J. (1972). The Foundations of Statistics, 2nd ed. Dover, New York. · Zbl 0276.62006
[25]	Sebastiani, P. (1996a). Discussion of the papers by Atkinson (1996) and by Bates et al. (1996). J. Roy. Statist. Soc. Ser. B 58 96-97. Sebastiani, P. (1996b). On the derivatives of matrix powers. SIAM J. Matrix Anal. Appl. 17 640-648.
[26]	Silvey, S. D. (1980). Optimal Design. Chapman and Hall, London. · Zbl 0468.62070
[27]	Stone, M. (1959). Application of a measure of information to the design and comparison of regression experiments. Ann. Math. Statist. 30 55-70. · Zbl 0094.13602 · doi:10.1214/aoms/1177706359
[28]	Verdinelli, I. (1996). Bayesian design for the normal linear model with unknown error variance. Technical Report 647, Carnegie Mellon Univ. · Zbl 0974.62028
[29]	Verdinelli, I. and Kadane, J. B. (1992). Bayesian design for maximizing information and outcome. J. Amer. Statist. Assoc. 87 510-515. JSTOR: · doi:10.2307/2290284
[30]	Whittle, P. (1973). Some general points on the theory of optimal experimental design. J. Roy. Statist. Soc. Ser. B 35 123-130. JSTOR: · Zbl 0282.62065
[31]	Zacks, S. (1977). Problems and approaches in design of experiments for estimation and testing in nonlinear models. In Multivariate Analysis IV (P. R. Krishnaiah, ed.). North-Holland, Amsterdam. · Zbl 0368.62052