EstimaÃÃo pontual em regressÃo Beta: aspectos computacionais

AUTOR(ES)
DATA DE PUBLICAÇÃO

2007

RESUMO

We consider the beta regression model proposed by Ferrari &Cribari-Neto (2004), which is tailored to situations where the response is restricted to the standard unit interval and has a regression structure involving regressors and unknown parameters. Our chief interest is the evaluation of several nonlinear optimization methods in the context of maximizing the beta regression log-likelihood function. The numerical evidence from Monte Carlo simulations and empirical analyses based on real data favors the Newton and BFGS algorithms, which are fast, accurate and behave well even in unfavorable situations such as the existence of leverage points and high correlation amongst regressors

ASSUNTO(S)

beta regression mÃxima verossimilhanÃa regressÃo beta maximum likelihood estatistica otimizaÃÃo optimization

Documentos Relacionados