Abstract
Researchers often find that nonlinear regression models are more applicable for modeling various biological, physical, and chemical processes than the linear ones since they tend to fit the data well and since these models (and model parameters) are more scientifically meaningful. These researchers are thus often in a position of requiring optimal or near-optimal designs for a given nonlinear model. A common shortcoming of most optimal designs for nonlinear models used in practical settings, however, is that these designs typically focus only on (first-order) parameter variance or predicted variance, and thus ignore the inherent nonlinear of the assumed model function. Another shortcoming of optimal designs is that they often have only p support points, where p is the number of model parameters. Measures of marginal curvature, first introduced in Clarke (1987) and further developed in Haines et al. (2004), provide a useful means of assessing this nonlinearity. Other relevant developments are the second-order volume design criterion introduced in Hamilton and Watts (1985) and extended in O’Brien (1992) and O’Brien et al. (2010), and the second-order MSE criterion developed and illustrated in Clarke and Haines (1995). This chapter underscores and highlights various robust design criteria and those based on second-order (curvature) considerations. These techniques, easily coded in the R and SAS/IML software packages, are illustrated here with several key examples.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Atkinson, A.C.: Planning experiments to detect inadequate regression models. Biometrika 59, 275–293 (1972)
Atkinson, A.C., Donev, A.N., Tobias, R.D.: Optimum Experimental Designs, with SAS. Oxford, New York (2007)
Bates, D.M., Watts, D.G.: Nonlinear Regression Analysis and Its Applications. Wiley, New York (1988)
Beale, E.M.L.: Confidence regions in non-linear estimation (with Discussion). J. R. Stat. Soc. B 22, 41–88 (1960)
Clarke, G.P.Y.: Marginal curvatures and their usefulness in the analysis of nonlinear regression models. J. Am. Stat. Assoc. 82, 844–850 (1987)
Clarke, G.P.Y., Haines, L.M.: Optimal design for models incorporating the richards function. In: Seeber, G.U.H., Francis, B.J., Hatzinger, R., Steckel-Berger, G. (eds.) Statistical Modelling. Springer, New York (1995)
Cook, R.D., Witmer, J.A.: A note on parameter-effects curvature. J. Am. Stat. Assoc. 80, 872–878 (1985)
Dette, H., O’Brien, T.E.: Optimality criteria for regression models based on predicted variance. Biometrika 86, 93–106 (1999)
Govaerts, B.: Discussion of the papers by Atkinson, and Bates et al. J. R. Stat. Soc. B 58, 95–111 (1996)
Haines, L.M., O’Brien, T.E., Clarke, G.P.Y.: Kurtosis and curvature measures for nonlinear regression models. Stat. Sinica 14, 547–570 (2004)
Hamilton, D.C., Watts, D.G.: A quadratic design criterion for precise estimation in nonlinear regression models. Technometrics 27, 241–250 (1985)
Kiefer, J.: General equivalence theory for optimal designs (approximation theory). Ann. Stat. 2, 849–879 (1974)
Kiefer, J., Wolfowitz, J.: The equivalence of two extremum problems. Can. J. Math. 12, 363–366 (1960)
O’Brien, T.E.: A note on quadratic designs for nonlinear regression models. Biometrika 79, 847–849 (1992)
O’Brien, T.E.: A new robust design strategy for sigmoidal models based on model nesting. In: Dutter, R., Grossmann, W. (eds.) CompStat 1994, pp. 97–102. Physica-Verlag, Heidelberg (1994)
O’Brien, T.E.: Optimal design and lack of fit in nonlinear regression models. In: Seeber, G.U.H., Francis, B.J., Hatzinger, R., Steckel-Berger, G. (eds.) Statistical Modelling, pp. 201–206. Springer, New York (1995)
O’Brien, T.E.: Efficient experimental design strategies in toxicology and bioassay. Stat. Optim. Inf. Comput. 4, 99–106 (2016)
O’Brien, T.E., Funk, G.M.: A gentle introduction to optimal design for regression models. Am. Stat. 57, 265–267 (2003)
O’Brien, T.E., Rawlings, J.O.: A non-sequential design procedure for parameter estimation and model discrimination in nonlinear regression models. J. Stat. Plann. Infer. 55, 77–93 (1996)
O’Brien, T.E., Chooprateep, S., Homkham, N.: Efficient geometric and uniform design strategies for sigmoidal regression models. S. Afr. Stat. J. 43, 49–83 (2009)
O’Brien, T.E., Jamroenpinyo, S., Bumrungsup, C.: Curvature measures for nonlinear regression models using continuous designs with applications to optimal design. Involve J. Math. 3, 317–332 (2010)
Pawitan, Y.: In All Likelihood: Statistical Modelling and Inference Using Likelihood. Oxford University Press, Oxford (2013)
Pukelsheim, F.: Optimal Design of Experiments. Wiley, New York (1993)
Pukelsheim, F., Rieder, S.: Efficient rounding of approximate designs. Biometrika 79, 763–770 (1992)
Ratkowsky, D.A.: Nonlinear Regression Modeling: A Unified Practical Approach. Marcel Dekker, New York (1983)
Seber, G.A.F., Wild, C.J.: Nonlinear Regression. Wiley, New York (1989)
Silvey, S.D.: Optimal Design: An Introduction to the Theory for Parameter Estimation. Chapman and Hall, London (1980)
Studden, W.J.: Some robust-type D-optimal designs in polynomial regression. J. Am. Stat. Assoc. 77, 916–921 (1982)
White, L.V.: An extension of the general equivalence theorem to nonlinear models. Biometrika 60, 345–348 (1973)
Acknowledgements
The author expresses his appreciation to the J. William Fulbright Foreign Scholarship Board for ongoing grant support and to Chiang Mai University (Thailand), Gadjah Mada University and Islamic University (Indonesia), Kathmandu University (Nepal), and Vietnam National University (Hanoi) for kind hospitality during recent research visits.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2018 Springer International Publishing AG
About this paper
Cite this paper
O’Brien, T.E. (2018). Contemporary Robust Optimal Design Strategies. In: Tez, M., von Rosen, D. (eds) Trends and Perspectives in Linear Statistical Inference . Contributions to Statistics. Springer, Cham. https://doi.org/10.1007/978-3-319-73241-1_11
Download citation
DOI: https://doi.org/10.1007/978-3-319-73241-1_11
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-73240-4
Online ISBN: 978-3-319-73241-1
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)
Keywords
- Binary logistic model
- Experimental design
- Generalized nonlinear modeling
- Goodness-of-fit
- Lack-of-fit
- Robustness