Article Sidebar

Submitted
May 19, 2017
Accepted
September 4, 2017
Published
January 21, 2018
Download
PDF
Statistic
Read Counter : 856 Download : 493

Downloads

Download data is not yet available.

Main Article Content

Abstract

Random parameter models have been found to outperform xed pa-rameter models to estimate dose-response relationships with independent errors. Amajor restriction, however, is that the responses are assumed to be normally andsymmetrically distributed. The purpose of this paper is to analyze Bayesian infer-ence of random parameter response models in the case of independent responseswith normal and skewed, heavy-tailed distributions by way of Monte Carlo simu-lation. Three types of Bayesian estimators are considered: one applying a normal,symmetrical prior distribution, a second applying a Skew-normal prior and, a thirdapplying a Skew-t-distribution. We use the relative bias (RelBias) and Root MeanSquared Error (RMSE) as valuation criteria. We consider the commonly applied lin-ear Quadratic and the nonlinear Spillman-Mitscherlich dose-response models. Onesimulation examines the performance of the estimators in the case of independent,normally and symmetrically distributed responses; the other in the case of indepen-dent responses following a heavy-tailed, Skew-t-distribution. The main nding isthat the estimator based on the Skew-t prior outperforms the alternative estima-tors applying the normal and Skew-normal prior for skewed, heavy-tailed data. Fornormal data, the Skew-t prior performs approximately equally well as the Skew-normal and the normal prior. Furthermore, it is more ecient than its alternatives.Overall, the Skew-t prior seems to be preferable to the normal and Skew-normal fordose-response modeling.

Keywords

Dose-response model Bayesian estimation Gibbs sampler Random parameter model Skew-normal distribution Skew-t distribution.

Article Details

Author Biographies

Mohammad Masjkur, Bogor Agricultural University

Department of Statistics, Faculty of Mathematics and Natural Sciences

Henk Folmer, University of Groningen

Faculty of Spatial Sciences
How to Cite
Masjkur, M., & Folmer, H. (2018). Bayesian Estimation of Random Parameter Models of Responses with Normal and Skew-t Distibutions Evidence from Monte Carlo Simulation. Journal of the Indonesian Mathematical Society, 24(1), 27–50. https://doi.org/10.22342/jims.24.1.516.27-50
Download Citation
Endnote/Zotero/Mendeley (RIS)
BibTeX
Crossref
Scopus
Google Scholar
Europe PMC

References

  1. Arellano-Valle, R. B., Bolfarine, H., and Lachos, V. H., Skew-normal linear mixed models,
  2. J. Data Sci., 3 (2005), 415-438.
  3. Arellano-Valle, R. B., Bolfarine, H., and Lachos, V. H., Bayesian inference for Skew-normal
  4. linear mixed models, J. Appl. Stat., 3 (2007), 663-682.
  5. Bandyopadhyay, D., Lachos, V. H., Castro, L. M., and Dey, D., "Skew-normal/independent
  6. linear mixed models for censored responses with applications to HIV viral loads", Biom. J.
  7. (2012), 405-425.
  8. Bandyopadhyay, D., Castro, L. M., Lachos, V. H., and Pinheiro, H. P., "Robust joint non-
  9. linear mixed-eects models and diagnostics for censored HIV viral loads with CD4 measure-
  10. ment error", J. Agric. Biol. Environ. Stat. 20 (2015), 121-139.
  11. Boyer, C.N., Larson, J.A., Roberts, R.K., McClure, A.T., Tyler, D.D., and Zhou, V., Sto-
  12. chastic corn yield response functions to nitrogen for corn after corn, corn after cotton, and
  13. corn after soybeans, J. Agric. Appl. Econ., 45 (2013), 669-681.
  14. Brorsen, B.W., Using Bayesian estimation and decision theory to determine the optimal level
  15. of nitrogen in cotton, Selected Paper, Southern Agricultural Economics Association, Orlando,
  16. Florida, (2013).
  17. de Souza, F. A., Malheiros, E. B., and Carneiro, P. R. O., Positioning and number of nutri-
  18. tional levels in dose-response trials to estimate the optimal-level and the adjustment of the
  19. models, Cincia Rural, Santa Maria, 44 (2014), 1204-1209.
  20. Gelman, A., Carlin J. B., Stern H. S., Dunson D. B., Vehtari A, and Rubin, D.B., Bayesian
  21. Data Analysis, Chapman Hall/CRC, New York, 2014.
  22. Hagenbuch, N., A comparison of four methods to analyse a non-linear mixed-eects model
  23. using simulated pharmacokinetic data, Master Thesis, Department of Mathematics, Swiss
  24. Federal Institute of Technology Zurich, 2011.
  25. Harring, J. R. and Liu J., A comparison of estimation methods for nonlinear mixed eects
  26. models under model misspecication and data sparseness: a simulation study, Journal of
  27. Modern Applied Statistical Methods, 15 (2016), 539-569.
  28. Jara, A., Quintana, F., San Martin, E., Linear mixed models with skew-elliptical distribu-
  29. tions: a Bayesian approach, Comput. Statist. Data Anal., 52 (2008), 5033-5045.
  30. Kery, M., Introduction to WinBUGS for Ecologists: A Bayesian approach to regression,
  31. ANOVA, mixed models and related analyses, Academic Press, Amsterdam, The Netherlands,
  33. Lachos, V. H., Dey, D. K., Cancho, V. G., Robust linear mixed models with skew-normal
  34. independent distributions from a Bayesian perspective, J. Statist. Plann. Inference, 139
  35. (2009), 4098-4110.
  36. Lachos, V. H., Ghosh P., and Arellano-Valle, R. B., Likelihood based inference for skew-
  37. normal independent linear mixed models, Statist. Sinica., 20 (2010), 303-322.
  38. Lachos, V. H., Castro L. M., and Dey D. K., Bayesian inference in nonlinear mixed-eects
  39. models using normal independent distributions, Comput. Statist. Data Anal., 64 (2013),
  40. -252.
  41. Lange, K. and Sinsheimer, J., Normal/independent distributions and their applications in
  42. robust regression, J. Comput. Graph. Stat., 2 (1993), 175-198.
  43. Lopez-Bellido, R.J., Castillo, J. E., and Lopez-Bellido, L., Comparative response of bread
  44. and durum wheat cultivars to nitrogen fertilizer in a rainfed Mediterranean environment: soil
  45. nitrate and N uptake and eciency, Nutr. Cycl. Agroecosyst., 80 (2008), 121-130.
  46. Makowski, D., Wallach, D., and Meynard, J.M., Statistical methods for predicting the re-
  47. sponses to applied N and for calculating optimal N rates, Agron. J., 93 (2001), 531-539.
  48. Makowski, D. and Wallach, D., It pays to base parameter estimation on a realistic description
  49. of model errors, Agronomie, 22 (2002), 179-189.
  50. Makowski, D. and Lavielle, M., Using SAEM to estimate parameters of response to applied
  51. fertilizer, J. Agric. Biol. Environ. Stat., 11 (2006), 45-60.
  52. Ouedraogo, F. B. and Brorsen, B. W., Bayesian estimation of optimal nitrogen rates with a
  53. non-normally distributed stochastic plateau function, The Southern Agricultural Economics
  54. Association (SAEA) Annual Meeting, Dallas, Texas, 2014.
  55. Park, S.C., Brorsen, B.W., Stoecker, A.L. and Hattey, J.A., Forage response to swine euent:
  56. a Cox nonnested test of alternative functional forms using a fast double bootstrap, J. Agr.
  57. Appl. Econ., 44 (2012), 593-606.
  58. Pinheiro, J., Bornkamp, B., Glimm, E., and Bretz, F., Model-based dose nding under model
  59. uncertainty using general parametric models, Stat. Med., 33 (2014), 1646-1661.
  60. Plan, E. L., Maloney, A., Mentr, F., Karlsson, M. O., and Bertrand J., Performance compar-
  61. ison of various maximum likelihood nonlinear mixed-eects estimation methods for dosere-
  62. sponse models, The AAPS J., 14 (2012), 420-432.
  63. Plummer, M., JAGS: A program for analysis of Bayesian graphical models using Gibbs sam-
  64. pling, Proceedings of the 3rd International Workshop on Distributed Statistical Computing,
  65. Vienna, Austria, 2003.
  66. Rizzo, M. L., Statistical Computing with R, Chapman Hall/CRC, New York, 2008.
  67. Rosa, G. J. M., Padovani, C. R., and Gianola, D., Robust linear mixed models with nor-
  68. mal/independent distributions and Bayesian MCMC implementation, Biom. J., 45 (2003),
  69. -590.
  70. Sain, G. E., and Jauregui, M.A., Deriving fertilizer recommendations with a
  71. exible functional
  72. form, Agron. J., 85 (1993), 934-937.
  73. Spiegelhalter, D. J., Best, N. G., Carlin, B. P., and van der Linde, A., Bayesian measures of
  74. model complexity and t, J. R. Stat. Soc. Ser. B. Stat. Methodol., 64 (2002), 583-639.
  75. Spiegelhalter, D. J., Best, N. G., Carlin, B. P., and van der Linde, A., The deviance infor-
  76. mation criterion: 12 years on, J. R. Stat. Soc. Ser. B. Stat. Methodol., 76 (2014), 485-493.
  77. Su, Y.S., and Yajima, M., R2jags: A package for running jags from R, R package version
  78. 5-7, 2015.
  79. Tanner, M. A., Tools for Statistical Inference: Methods for the Exploration of Posterior
  80. Distributions and Likelihood Functions, Springer-Verlag New York, 1993.
  81. Tembo, G., Brorsen, B. W., Epplin, F. M., and Tostao, E., Crop input response functions
  82. with stochastic plateaus, Amer. J. Agr. Econ., 90 (2008), 424-434.
  83. Tumusiime, E., Brorsen, B.W., Mosali, J., Johnson, J., Locke, J. and Biermacher, J.T.,
  84. Determining optimal levels of nitrogen fertilizer using random parameter models, J. Agr.
  85. Appl. Econ., 43 (2011), 541-552.
  86. Wallach, D., Regional optimization of fertilization using a hierarchical linear model, Biomet-
  87. rics, 51 (1995), 338-346.
  88. Ward, E. J., A review and comparison of four commonly used Bayesian and maximum like-
  89. lihood model selection tools, Ecological Modelling, 211 (2008), 110.
  90. WHO, Principles for Modelling Dose-Response For The Risk Assessment of Chemicals,
  91. WHO Press, World Health Organization, Geneva, Switzerland, 2009.

Similar Articles

You may also start an advanced similarity search for this article.