Semiparametric Smoothing Spline in Joint Mean and Dispersion Models with Responses from the Biparametric Exponential Family: A Bayesian Perspective

  • Héctor Zárate Department of Statistics, Universidad Nacional de Colombia, Bogot, Colombia
  • Edilberto Cepeda Department of Statistics, Universidad Nacional de Colombia, Bogot, Colombia
Keywords: Markov chain Monte Carlo, Double generalized linear mixed models, Biparametric exponential family, Spline models


This article extends the fusion among various statistical methods to estimate the mean and variance functions in heteroscedastic semiparametric models when the response variable comes from a two-parameter exponential family distribution. We rely on the natural connection among smoothing methods that use basis functions with penalization, mixed models and a Bayesian Markov Chain sampling simulation methodology. The significance and implications of our strategy lies in its potential to contribute to a simple and unified computational methodology that takes into account the factors that affect the variability in the responses, which in turn is important for an efficient estimation and correct inference of mean parameters without the requirement of fully parametric models. An extensive simulation study investigates the performance of the estimates. Finally, an application using the Light Detection and Ranging technique, LIDAR, data highlights the merits of our approach.


R. Littell and O. Schabenberger, SAS for Mixed Models. No. 2, 2006.

J. Pinheiro and D. Bates, Mixed-Effects Models in S and S-Plus. Springer Verlag, 2009.

D. Spiegelhalter and N. Best, “Bayesian approaches to multiple sources of evidence in complex cost-effectiveness modelling,” Statistics in Medicine, no. 23, pp. 3687 – 3709, 2003.

M. B. Denison, D and F. Smith, “A bayesian cart algorithm,” Biometrika, no. 2, pp. 363 – 367, 1998.

D. Nott, “Semiparametric estimation of mean and variance functions for non-gaussian data,” Computational Statistics, no. 3-4, pp. 603–620, 2006.

I. Gijbels and I. Prosdocimi, “Flexible mean and dispersion function estimation in extended generalized additive models,” Communications in statistics - Theory and Methods, no. 41, pp. 3259 – 3277, 2012.

D. Ruppert, M. Wand, and R. J. Carroll, “Semiparametric regression during 2003-2007,” Electronic Journal of Statistics, vol. 3, pp. 1193–1256, 2009.

C. Crainiceanu, “Spatially adaptative bayesian penalized splines with heteroscedastic errors,” Journal of Computational and Graphical Statistics, no. 2, pp. 265–288, 2007.

D. Xu and Z. Zhang, “A semiparametric bayesian approach to joint mean and variance models,” Statistics & Probability Letters, vol. 83, no. 7, pp. 1624 – 1631, 2013.

M. Mencitas and M. Wand, “Variational inference for heteroscedastic semiparametric regression,” School of mathematical sciences, University of Technology. Sydney, Australia, 2014.

E. Cepeda and D. Gamerman, “Bayesian modeling of variance heterogeneity in normal regression models,” J. Prob.Stat, vol. 14, pp. 207–221, 2001.

Y. Ma and C. R. J., “Locally efficient estimators for semiparametric models with measurement error,” Journal of the American Statistical Association, no. 101, p. 1465-1474, 2006.

C. R. R. D. Berry, S., “Bayesian smoothing and regression splines for measurement error problems,” Journal of the American Statistical Association, no. 457, pp. 160–169, 2011.

M. B. Eilers, P. and M. Durbn, “Twenty years of p-splines,” SORT (Statistics and Operations Research Transactions), no. 39, 2014.

N. N. N. D. Tran, M. and R. Kohn, “Bayesian deep net glm and glmm,” SORT ( arXiv:1805.10157v1 [stat.CO]), 2018.

E. V. T. N. Robert, C. P. and W. C., “Accelerating mcmc algorithms,” Journal of the American Statistical Association, 2018.

B. M. Currie, I., “Flexible smoothing with p-splines : a unified approach,” Statistical Modelling, no. 4, pp. 333–349, 2002.

C. Gu, Smoothing Spline ANOVA Models. Springer, West Lafayette,USA, 2002.

G. Wahba, Spline Models for Observational Data. Society for Industrial and Applied Mathematics, 1990.

P. Green and B. Silverman, Nonparametric Regression and Generalized Linear Models: A roughness penalty approach. Chapman and Hall, London, 1994.

S. C. T. C. C. R. Nosedal-Sanchez, A., “Reproducing kernel hilbert spaces for penalized regression : A tutorial,” The American Statistician, no. 66, pp. 50–60, 2012.

W. D. Pierce, N., “Penalized splines and reproducible kernel methods,” American Statistical Association, no. 3, pp. 233–240, 2006.

D. K. Dey, A. E. Gelfand, and F. Peng, “Overdispersed generalized linear models,” Journal of statistical planning and inference,, vol. 64, no. 64, pp. 93–108, 1997.

E. Cepeda, Variability modeling in Generalized Linear models. PhD thesis, Unpublished Ph.D thesis, Matematics Institute Universidade Federal do Rio de Janeiro, 2001.

D. Gamerman, “Sampling from the posterior distribution in generalized linear mixed models,” Instituto de matemtica, Universidade Federal do Rio de Janeiro, pp. 59 – 68, 1997.

E. Cepeda, J. A. Achcar, and L. G. Lopera, “Bivariate beta regression models: joint modeling of the mean, dispersion and association parameters,” Journal of Applied statistics, vol. 41, pp. 677–687, Marzo 2014.

E. Cepeda and D. Gamerman, “Bayesian methodology for modeling parameters in the two parametric exponential family,” Estadstica, vol. 57, pp. 93–105, 2005.

D. Ruppert, M.Wand, U. Holst, and O. Hssjer, “Local polynomial variance-function estimation,” Technometrics, no. 39, pp. 262–273, 1997.

L. Chelsea, “Mcmc in sas: From scratch or by proc,” Wetern users of SAS software 2016, vol. 1, no. 1, pp. 1 – 19, 2016.

How to Cite
Zárate, H., & Cepeda, E. (2021). Semiparametric Smoothing Spline in Joint Mean and Dispersion Models with Responses from the Biparametric Exponential Family: A Bayesian Perspective. Statistics, Optimization & Information Computing, 9(2), 351-367.
Research Articles