Review Articles

Locally R-optimal designs for a class of nonlinear multiple regression models

Lei He ,

Department of Statistics, Anhui Normal University, Wuhu, People's Republic of China

Rong-Xian Yue

Department of Mathematics, Shanghai Normal University, Shanghai, People's Republic of China

yue2@shnu.edu.cn

Pages | Received 09 Jun. 2022, Accepted 27 Nov. 2022, Published online: 12 Dec. 2022,
  • Abstract
  • Full Article
  • References
  • Citations

This paper concerns with optimal designs for a wide class of nonlinear models with information driven by the linear predictor. The aim of this study is to generate an R-optimal design which minimizes the product of the main diagonal entries of the inverse of the Fisher information matrix at certain values of the parameters. An equivalence theorem for the locally R-optimal designs is provided in terms of the intensity function. Analytic solutions for the locally saturated R-optimal designs are derived for the models having linear predictors with and without intercept, respectively. The particle swarm optimization method has been employed to generate locally non-saturated R-optimal designs. Numerical examples are presented for illustration of the locally R-optimal designs for Poisson regression models and proportional hazards regression models.

References

  • Atkinson, A. C., Donev, A. N., & Tobias, R. D. (2007). Optimum experimental designs, with SAS. Oxford University Press. 
  • Chernoff, H. (1953). Locally optimal designs for estimating parameters. The Annals of Mathematical Statistics24(4), 586–602. https://doi.org/10.1214/aoms/1177728915 
  • Dette, H. (1997). Designing experiments with respect to ‘standardized’ optimality criteria. Journal of the Royal Statistical Society: Series B (Statistical Methodology)59(1), 97–110. https://doi.org/10.1111/rssb.1997.59.issue-1 
  • Fedorov, V. V. (1972). Theory of optimal experiments. Academic Press. 
  • Freise, F., Gaffke, N., & Schwabe, R. (2021). The adaptive Wynn algorithm in generalized linear models with univariate response. The Annals of Statistics49(2), 702–722. https://doi.org/10.1214/20-AOS1974
  • Hao, H., Zhu, X., Zhang, X., & Zhang, C. (2021). R-optimal design of the second-order Scheffé mixture model. Statistics & Probability Letters173(C), 109069. https://doi.org/10.1016/j.spl.2021.109069 
  • Idais, O., & Schwabe, R. (2021). Analytic solutions for locally optimal designs for gamma models having linear predictors without intercept. Metrika84(1), 1–26. https://doi.org/10.1007/s00184-019-00760-3 
  • Kalish, L. A., & Harrington, D. P. (1988). Efficiency of balanced treatment allocation for survival analysis. Biometrics44(3), 815–821. https://doi.org/10.2307/2531593 
  • Konstantinou, M., Biedermann, S., & Kimber, A. (2014). Optimal designs for two-parameter nonlinear models with application to survival models. Statistica Sinica24(1), 415–428. https://doi.org/10.5705/ss.2011.271 
  • Li, K. H., Lau, T. S., & Zhang, C. (2005). A note on D-optimal designs for models with and without an intercept. Statistical Papers46(3),451–458. https://doi.org/10.1007/BF02762844 
  • Liu, P., Gao, L. L., & Zhou, J. (2022). R-optimal designs for multi-response regression models with multi-factors. Communications in Statistics – Theory and Methods51(2), 340–355. https://doi.org/10.1080/03610926.2020.1748655 
  • Liu, X., & Yue, R.-X. (2020). Elfving's theorem for R-optimality of experimental designs. Metrika83(4), 485–498. https://doi.org/10.1007/s00184-019-00728-3 
  • Pukelsheim, F., & Torsney, B. (1993). Optimal weights for experimental designs on linearly independent support points. The Annals of Statistics19(3), 1614–1625. https://doi.org/10.2307/2241966
  • Radloff, M., & Schwabe, R. (2019). Locally D-optimal designs for non-linear models on the k-dimensional ball. Journal of Statistical Planning and Inference203, 106–116. https://doi.org/10.1016/j.jspi.2019.03.004
  • Russell, K. G., Woods, D. C., Lewis, S. M., & Eccleston, E. C. (2009). D-optimal designs for Poisson regression models. Statistica Sinica19(2), 721–730. https://doi.org/10.2307/24308852 
  • Schmidt, D. (2019). Characterization of c-, L- and ϕkϕk-optimal designs for a class of non-linear multiple-regression models. Journal of the Royal Statistical Society: Series B (Statistical Methodology)81(1), 101–120. https://doi.org/10.1111/rssb.12292 
  • Schmidt, D., & Schwabe, R. (2015). On optimal designs for censored data. Metrika78(3), 237–257. https://doi.org/10.1007/s00184-014-0500-1 
  • Schmidt, D., & Schwabe, R. (2017). Optimal design for multiple regression with information driven by the linear predictor. Statistica Sinica27(3), 1371–1384. https://doi.org/10.5705/ss.202015.0385
  • Silvey, S. D. (1980). Optimal design. Chapman and Hall. 
  • Whittle, P. (1973). Some general points in the theory of optimal experimental designs. Journal of the Royal Statistical Society: Series B (Methodological)35(1), 123–130. https://doi.org/10.1111/j.2517-6161.1973.tb00944.x 
  • Yang, M., Biedermann, S., & Tang, E. (2013). On optimal designs for nonlinear models: A general and efficient algorithm. Journal of the American Statistical Association108(504), 1411–1420. https://doi.org/10.1080/01621459.2013.806268 

To cite this article: Lei He & Rong-Xian Yue (2023) Locally R-optimal designs for a class of nonlinear multiple regression models, Statistical Theory and Related Fields, 7:2, 107-120, DOI: 10.1080/24754269.2022.2153540

To link to this article: https://doi.org/10.1080/24754269.2022.2153540