Estimation and diagnostic in Birnbaum-Saunders skew-normal models / Estimação e diagnostico em modelos Birnbaum-Saunders skew-normal
AUTOR(ES)
Lucia Rolim Santana
DATA DE PUBLICAÇÃO
2009
RESUMO
The class of models Birnbaum-Saunders (BS) appeared in problems of fatigue of materials (which is a structural damage that occurs when a material is exposed to stress and tension). Recently, this model has been applied in areas outside the context of fatigue of materials and engineering, for example in health sciences, environmental, forestry, demographic, actuarial, financial, among others. As the BS distribution has the property to describe cumulative degradation processes. In this work we present a study of the BS model based on Skew-normal distribution. As a byproduct consider the model of linear regression log-Birnbaum-Saunders (log-BS). To obtain estimates of maximum likelihood we use the EM algorithm. Furthermore, we present a study of the analysis of global and local influence, through the method of Zhu and Lee (2001) to incomplete data. Illustrate the proposed methodology with data found in literature
ASSUNTO(S)
influencia local expectation-mazimization algorithms local influence analise de regressão - modelos birnbaum-saunders analise de regressão algoritmos de expectativa de maximização regression anlaysis
ACESSO AO ARTIGO
http://libdigi.unicamp.br/document/?code=000445637Documentos Relacionados
- Modelos não lineares sob a classe de distribuições misturas da escala skew-normal
- Mixtures modelling using scale mixtures of skew-normal distribution
- Avaliação dos desempenhos de estimadores para os parâmetros da distribuição Birnbaum-Saunders
- InferÃncia sobre os parÃmetros da distribuiÃÃo Birnbaum-Saunders bi-paramÃtrica
- Utilização da função Birnbaum-Saunders para modelagem da distribuição diamétrica de povoamentos equiâneos de eucalipto