zbMATH — the first resource for mathematics

Likelihood analysis for a class of simplex mixed models. (English) Zbl 1449.62154
Summary: This paper describes the specification, estimation and comparison of simplex mixed models based on the likelihood paradigm. This class of models is suitable to deal with restricted response variables, such as rates, percentages, indexes and proportions. The estimation of simplex mixed models is challenged by the intractable integral in the likelihood function. We compare results obtained with three numerical integration methods Laplace, Gauss-Hermite and Quasi-Monte Carlo to solve such integral. The specification of simplex mixed models includes the choice of a link function for which we compare models fitted with logit, probit, complement log-log and Cauchy link functions. Furthermore, results from the simplex mixed models fitted to two datasets are compared with fits of beta, linear and non-linear mixed effects models. The first is a study concerning life quality of industry workers with data collected according to a hierarchical sampling scheme. The second corresponds to water quality measurements taken at 16 operating hydroelectric power plants in Paran\'a State, Brazil. Our results showed that the simplex mixed models provide the best fit between the approaches considered for the two datasets analyzed. None of the choices of the link function outperformed the others. Simulation studies were designed to check the properties of the maximum likelihood estimators and the computational implementation. The Laplace method provides the best balance between computational complexity and accuracy. The data sets and R code are available in the supplementary material.
62J05 Linear regression; mixed models
62P12 Applications of statistics to environmental and related topics
Full Text: Link
[1] Barndorff-Nielsen. O.E., and Jørgensen, B. (1991). Some parametric models on the simplex. Journal of Multivariate Analysis 39, 106-116. · Zbl 0739.62015
[2] Bolker, B., and R Core Team (2014).bbmle: Tools for general maximum likelihood estimation. R package version 1.0.17. URL:CRAN.R-project.org/package=bbmle
[3] Bonat, W.H., Ribeiro Jr, P.J., and Zeviani, W.M. (2012). Regression models with response on the unit interval: Specification, estimation and comparison. Biometric Brazilian Journal 30, 415-431.
[4] Bonat, W.H., Ribeiro Jr, P.J., and Zeviani, W.M. (2015). Likelihood analysis for a class of beta mixed models. Journal of Applied Statistics 42, 252-266.
[5] Bonat, W.H., Ribeiro Jr, P.J., and Shimakura, S.E. (2015). Bayesian analysis for a class of beta mixed models. Chilean Journal of Statistics 6, 3-13. · Zbl 1449.62057
[6] Bonat, W.H., and Ribeiro Jr, P.J. (2016). Practical likelihood analysis for spatial generalized linear mixed models. Environmetrics 27, 83-89.
[7] Cepeda, E. (2001). Variability Modeling in Generalized Linear Models. Unpublished Ph.D. Thesis. Mathematics Institute, Universidade Federal do Rio de Janeiro, Brazil.
[8] Ferrari, S.L.P, and Cribari-Neto, F. (2004). Beta regression for modelling rates and proportions. Journal of Applied Statistics 31, 799-815. · Zbl 1121.62367
[9] Figueroa-Z´u˜niga, J.I. and Arellano-Valle, R.B., and Ferrari, S.L.P. Mixed beta regression: A Bayesian perspective. Computational Statistics and Data Analysis 61, 137-147. · Zbl 1348.62194
[10] Gilbert, P., and Varadhan, R. (2012).numDeriv: Accurate numerical derivatives. R package version 2012.9-1. URL:CRAN.R-project.org/package=numDeriv
[11] Jørgensen, B. (1997). The Theory of Dispersion Models. Chapman & Hall, London.
[12] Kieschnick, R., and McCullough, B.D. (2003). Regression analysis of variates observed on (0,1): percentages, proportions and fractions. Statistical Modelling 3, 193-213. · Zbl 1070.62056
[13] Lindfield, G.R., and Penny, J.E.T. (1989). Microcomputers in Numerical Analysis. Halsted Press, New York. · Zbl 0662.65001
[14] L´opez, F.O. (2013). A Bayesian approach to parameter estimation in simplex regression model: A comparison with beta regression. Revista Colombiana de Estad´ıstica 36, 1-21.
[15] Moral, R.A., Hinde, J., and Dem´etrio, C.G.B. (2017). Half-normal plots and overdispersed models in R: ThehnpPackage. Journal of Statistical Software 81, 123.doi: 10.18637/jss.v081.i10.
[16] Nelder, J.A., and Wedderburn, R.W. (1972). Generalized linear models. Journal of the Royal Statistical Society, Series A 135, 370-384.
[17] Noh, M. and Lee, Y. (2007). REML estimation for binary data in GLMMs. Journal of Multivariate Analysis 98, 896-915. · Zbl 1113.62087
[18] Pan, J., and Thompson, R. (2007). Quasi-Monte Carlo estimation in generalized linear mixed models. Computational Statistics & Data Analysis 51, 5765-5775. · Zbl 1445.62196
[19] Paolino, P. (2001). Maximum likelihood estimation of models with Beta-distributed dependent variables. Political Analysis 9, 325-346.
[20] Qiu, Z., Song, P.X.K., and Tan, M. (2008). Simplex mixed-effects models for longitudinal proportional data. Scandinavian Journal of Statistics 35, 577-596. · Zbl 1197.62031
[21] R Core Team (2015). R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. URL:www.R-project.org
[22] Smyth, G., Hu, Y., Dunn, P., and Phipson, B. (2013).statmod: Statistical Modeling. R package version 1.4.17. URL:CRAN.R-project.org/package=statmod
[23] Song, P.X.K. (2007). Correlated Data Analysis: Modeling, Analytics, and Applications. Springer, New York. · Zbl 1132.62002
[24] Tierney, L., and Kadane, J. (1986). Accurate approximations for posterior moments and marginal densities. Journal of the American Statistical Association 81, 82-86. · Zbl 0587.62067
[25] Wuertz, D. (2012).fOptions: Basics of Option Valuation. R package version 2160.82. URL: CRAN.R-project.org/package=fOptions
[26] Zhang,P.,Qiu,Z.,andShi,C.(2014).simplexreg:Regressionanalysisof proportionaldatausingsimplexdistribution.Rpackageversion1.0.URL: CRAN.R-project.org/package=simplexreg
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.