Bayesian A-Optimal Designs for Gamma and Poisson Regression Models: An Algorithmic Approach


103 / 97

Authors

  • Mahesh Kumar Panda Department of Statistics, Ravenshaw University, Cuttack
  • Tofan Kumar Biswal Department of Statistics, Central University of Odisha, Sunabeda
  • V.K. Gupta Formerly at ICAR-IASRI, New Delhi

https://doi.org/10.56093/JISAS.V79I2.2

Keywords:

Fisher information matrix; Bayesian A-optimal design; Gamma regression model; Poisson regression model; Prior information.

Abstract

Bayesian design approaches leverage prior information about unknown parameters to improve the efficiency of experimental designs. In this study, 
we propose two algorithms for deriving Bayesian A-optimal designs for Gamma and Poisson regression models involving two explanatory factors. 
The experimental region is assumed to be a unit rectangle. The present work focuses specifically on vertex-type designs, where the support points of 
the initial designs are restricted to the vertices of the experimental region. Initially, the Fisher information matrix is computed based on the support 
points of the proposed locally optimal designs. Subsequently, Bayesian A-optimal designs are determined by incorporating prior distributions-namely, 
Uniform, Normal, Beta, and Gamma-on the unknown parameters. The corresponding optimal weights for each design are computed using MATLAB.

Downloads

Download data is not yet available.

References

1. Abdelbasit, K. M., and Plackett, R. L. (1983). Experimental design for

2. binary data. Journal of the American Statistical Association, 78,

3. 90-98.

4. Abebe, H. T., Tan, F. E., Van Breukelen, G. J., Serroyen, J., and Berger,

5. M. P. (2014). On the choice of a prior for Bayesian D-optimal

6. designs for the logistic regression model with a single predictor.

7. Communications in Statistics-Simulation and Computation, 43,

8. 1811-1824.

9. Atkinson, A.C., and Woods, D.C. (2015). Designs for generalized linear

10. models. In Handbook of Design and Analysis of Experiments, 7,

11. 471-514.

12. Chaloner, K., and Larntz, K. (1989). Optimal Bayesian design applied

13. to logistic regression experiments. Journal of Statistical Planning

14. and Inference, 21, 191-208.

15. Chaloner, K., andVerdinelli, I. (1995). Bayesian experimental design: A

16. review. Statistical science, 10, 273-304.

17. Chernoff, H. (1953). Locally optimal designs for estimating parameters.

18. The Annals of Mathematical Statistics, 24, 586-602.

19. Fedorov, V. V., and Leonov, S. L. (2013). Optimal design for nonlinear

20. response models. CRC Press.

21. Finney, D.J. (1978).Statistical Method in Biological Assay. Charles

22. Griffin and Company Ltd, 3rd edition.

23. Gaffke, N., Idais, O., and Schwabe, R. (2019). Locally optimal designs

24. for gamma models. Journal of Statistical Planning and Inference,

25. 203, 199-214.

26. Goldburd, M., Khare, A., Tevet, D., and Guller, D. (2016). Generalized

27. linear models for insurance rating. Casualty Actuarial Society,

28. CAS Monographs Series, 7, 2016.

29. Holling, J.P. and Schwabe, R. (2011). Simultaneous confidence bands

30. and optimal design for logistic regression. Journal of Statistical

31. Planning and Inference, 141, 717-727.

32. Idais, O. (2021). On local optimality of vertex type designs in

33. generalized linear models. Statistical Papers, 62, 1871-1898.

34. Idais, O., and Schwabe, R. (2021). Analytic solutions for locally

35. optimal designs for gamma models having linear predictors

36. without intercept. Metrika, 84, 1-26.

37. Khuri, A.I., Mukherjee, B., Sinha, B.K., and Ghosh, M. (2006). Design

38. issues for generalized linear models: A review, Statistical Science,

39. 21, 376-399.

40. McCullagh P, Nelder J (1989) Generalized linear models, 2nd edition.

41. Chapman and Hall, London

42. Meeker, W.Q., and Hahn, G.J. (1977). Asymptotically optimum over

43. stress tests to estimate the survival probability at a condition with

44. a low expected failure probability. Technometrics, 19, 381-399.

45. Myers, R.H., and Montgomery, D.C. (1997). A tutorial on generalized

46. linear models. Journal of Quality Technology, 29, 274-291.

47. Nelder, J.A., and Wedderburn, R.W. (1972). Generalized linear models.

48. Journal of the Royal Statistical Society Series A: Statistics in

49. Society, 135, 370-384.

50. Panda, M.K. and Biswal, T.K. (2025a). A-optimal designs for two

51. variable logistic regressionmodel with restricted design space.

52. Manuscript submitted

53. Panda, M.K., and Biswal, T.K. (2025b). R-optimal Designs for Logistic

54. Regression Model in Two Variables. Statistics and Applications,

55. (New Series), 23, 1–12.

56. Panda, M.K., Biswal, T.K. and Gupta, V.K. (2025). R-optimal designs

57. for gamma regression model with two parameters. Statistics and

58. Applications, (New Series), 23, 33-53.

59. Ryan, L., Hidiroglou, M.A., and Hidiroglou, Y. (2015). Assessing the

60. impact of potentiallyinfluential observations in logistic regression.

61. Survey Methodology, 41, 91-98.

62. Silvey, S. D. (1980).Optimal Design. Chapman and Hall.

63. Walker, S.H., and Duncan, D.B. (1967). Estimation of the probability

64. of an event as a function of several independent variables.

65. Biometrika, 54, 167-179

Downloads

Published

2026-07-24

Issue

Section

Articles

How to Cite

Mahesh Kumar Panda, Tofan Kumar Biswal, & V.K. Gupta. (2026). Bayesian A-Optimal Designs for Gamma and Poisson Regression Models: An Algorithmic Approach. Journal of the Indian Society of Agricultural Statistics, 79(2), 63-78. https://doi.org/10.56093/JISAS.V79I2.2
Citation