The use of empirical Bayesian methods to fit hypsometric relationship with the Curtis model
 with constraint in the parameter

Monica Fabiana Bento MOREIRA[1]

Selene LOIBEL1

Cláudio Roberto THIERSCH[2]

Marinho Gomes de ANDRADE1

§    ABSTRACT: The objective of this paper is to propose the use of an empirical Bayesian methods for fitting hypsometric relationships (diameter and height) for clones of \emph{Eucalyptus sp.}, with Curtis model with biological constraint imposed to the parameters. The posterior probability density is obtained via simulation technique with Monte Carlo Markov Chain (MCMC) methods. The proposed approach was applied for five real data sets and the results were compared to the obtained by the minimum square method, the superiority of the Bayesian method is highlighted.

§    KEYWORDS: Hypsometric model, Curtis model; empirical Bayesian methods; simulation MCMC; non linear regression model.



[1] Departamento de Matemática Aplicada e Estatística, Instituto de Ciências Matemáticas e de Computação – ICMC, Universidade de São o Paulo – USP, Caixa Postal 668, CEP: 13560-970, São Carlos, SP, Brasil. E-mail: monica@icmc.usp.br / sloibel@icmc.usp.br / marinho@icmc.usp.br

[2] Biometria e Inventário Florestal, Votorantim Celulose Papel, Rod. Gal Euryale Jesus Zerbine, s/n - São Silvestre - Km: 84, CEP:12300-000, Jacareí, SP, Brasil. E-mail: claudio.thiersch@vcp.com.br