A log-extended Weibull regression model


Autoria(s): SILVA, Giovana O.; ORTEGA, Edwin M. M.; CORDEIRO, Gauss M.
Contribuinte(s)

UNIVERSIDADE DE SÃO PAULO

Data(s)

18/10/2012

18/10/2012

2009

Resumo

A bathtub-shaped failure rate function is very useful in survival analysis and reliability studies. The well-known lifetime distributions do not have this property. For the first time, we propose a location-scale regression model based on the logarithm of an extended Weibull distribution which has the ability to deal with bathtub-shaped failure rate functions. We use the method of maximum likelihood to estimate the model parameters and some inferential procedures are presented. We reanalyze a real data set under the new model and the log-modified Weibull regression model. We perform a model check based on martingale-type residuals and generated envelopes and the statistics AIC and BIC to select appropriate models. (C) 2009 Elsevier B.V. All rights reserved.

CNPq

CAPES

Identificador

COMPUTATIONAL STATISTICS & DATA ANALYSIS, v.53, n.12, p.4482-4489, 2009

0167-9473

http://producao.usp.br/handle/BDPI/18941

10.1016/j.csda.2009.07.004

http://dx.doi.org/10.1016/j.csda.2009.07.004

Idioma(s)

eng

Publicador

ELSEVIER SCIENCE BV

Relação

Computational Statistics & Data Analysis

Direitos

restrictedAccess

Copyright ELSEVIER SCIENCE BV

Palavras-Chave #FAILURE RATE-FUNCTION #CENSORED-DATA #RESIDUALS #SURVIVAL #Computer Science, Interdisciplinary Applications #Statistics & Probability
Tipo

article

original article

publishedVersion