Estimation in the Birnbaum-Saunders distribution based on scale-mixture of normals and the EM-algorithm

Scale mixtures of normal (SMN) distributions are used for modeling symmetric data. Members of this family have appealing properties such as robust estimates, easy number generation, and efficient computation of the ML estimates via the EM-algorithm. The Birnbaum-Saunders (BS) distribution is a posit...

Descripción completa

Detalles Bibliográficos
Autores: Balakrishnan, N., Leiva, Víctor, Sanhueza, Antonio, Vilca, Filidor
Tipo de recurso: artículo
Fecha de publicación:2009
País:España
Institución:Universitat Politècnica de Catalunya (UPC)
Repositorio:UPCommons. Portal del coneixement obert de la UPC
Idioma:inglés
OAI Identifier:oai:upcommons.upc.edu:2099/8949
Acceso en línea:https://hdl.handle.net/2099/8949
Access Level:acceso abierto
Palabra clave:Numerical analysis--Simulation methods
Stochastic differential equations--Numerical solutions
Distribution (Probability theory)
Birnbaum-Saunders distribution
EM-algorithm
Kurtosis
Maximum likelihood methods
Robust estimation
Scale mixtures of normal distributions
Anàlisi numèrica
Distribució (Teoria de la probabilitat)
Classificació AMS::65 Numerical analysis::65C Probabilistic methods, simulation and stochastic differential equations
Classificació AMS::60 Probability theory and stochastic processes::60E Distribution theory
Àrees temàtiques de la UPC::Matemàtiques i estadística::Anàlisi numèrica
Àrees temàtiques de la UPC::Matemàtiques i estadística::Estadística matemàtica
Descripción
Sumario:Scale mixtures of normal (SMN) distributions are used for modeling symmetric data. Members of this family have appealing properties such as robust estimates, easy number generation, and efficient computation of the ML estimates via the EM-algorithm. The Birnbaum-Saunders (BS) distribution is a positively skewed model that is related to the normal distribution and has received considerable attention. We introduce a type of BS distributions based on SMN models, produce a lifetime analysis, develop the EM-algorithm for ML estimation of parameters, and illustrate the obtained results with real data showing the robustness of the estimation procedure.