			Home About us Contact

Mixture Representation (mixture + representation)

Distribution by Scientific Domains

Mathematics and Statistics	100%

Selected Abstracts

On the application and extension of system signatures in engineering reliability

NAVAL RESEARCH LOGISTICS: AN INTERNATIONAL JOURNAL, Issue 4 2008
Jorge Navarro
Abstract Following a review of the basic ideas in structural reliability, including signature-based representation and preservation theorems for systems whose components have independent and identically distributed (i.i.d.) lifetimes, extensions that apply to the comparison of coherent systems of different sizes, and stochastic mixtures of them, are obtained. It is then shown that these results may be extended to vectors of exchangeable random lifetimes. In particular, for arbitrary systems of sizes m < n with exchangeable component lifetimes, it is shown that the distribution of an m -component system's lifetime can be written as a mixture of the distributions of k -out-of- n systems. When the system has n components, the vector of coefficients in this mixture representation is precisely the signature of the system defined in Samaniego, IEEE Trans Reliabil R,34 (1985) 69,72. These mixture representations are then used to obtain new stochastic ordering properties for coherent or mixed systems of different sizes. © 2008 Wiley Periodicals, Inc. Naval Research Logistics, 2008 [source]

Two-part regression models for longitudinal zero-inflated count data

THE CANADIAN JOURNAL OF STATISTICS, Issue 2 2010
Marco Alfò
Abstract Two-part models are quite well established in the economic literature, since they resemble accurately a principal-agent type model, where homogeneous, observable, counted outcomes are subject to a (prior, exogenous) selection choice. The first decision can be represented by a binary choice model, modeled using a probit or a logit link; the second can be analyzed through a truncated discrete distribution such as a truncated Poisson, negative binomial, and so on. Only recently, a particular attention has been devoted to the extension of two-part models to handle longitudinal data. The authors discuss a semi-parametric estimation method for dynamic two-part models and propose a comparison with other, well-established alternatives. Heterogeneity sources that influence the first level decision process, that is, the decision to use a certain service, are assumed to influence also the (truncated) distribution of the positive outcomes. Estimation is carried out through an EM algorithm without parametric assumptions on the random effects distribution. Furthermore, the authors investigate the extension of the finite mixture representation to allow for unobservable transition between components in each of these parts. The proposed models are discussed using empirical as well as simulated data. The Canadian Journal of Statistics 38: 197,216; 2010 © 2010 Statistical Society of Canada Les modèles en deux parties sont bien établis dans la littérature économique puisqu'ils sont très similaires à un modèle principal-agent pour lequel les résultats homogènes, observables et dénombrables sont sujets à un critère de sélection (exogène et a priori). La première décision est représentée à l'aide un modèle de choix binaire et une fonction de lien probit ou logit tandis que la seconde peut être analysée à l'aide d'une loi discrète tronquée telle que la loi de Poisson tronquée, la loi binomiale négative, etc. Depuis peu, une attention particulière a été portée à la généralisation du modèle en deux parties pour prendre en compte les données longitudinales. Les auteurs présentent une méthode d'estimation semi-paramétrique pour les modèles en deux parties dynamiques et ils les comparent avec d'autres modèles alternatifs bien connus. Les sources hétérogènes qui influencent le premier niveau du processus de décision, c'est-à-dire la décision d'utiliser un certain service, sont censées influencer aussi la distribution (tronquée) des résultats positifs. L'estimation est faite à l'aide de l'algorithme EM sans présupposés paramétriques sur la distribution des effets aléatoires. De plus, les auteurs considèrent une généralisation à une représentation en mélange fini afin de permettre une transition non observable entre les différentes composantes de chacune des parties. Une discussion est faite sur les modèles présentés en utilisant des données empiriques ou simulées. La revue canadienne de statistique 38: 197,216; 2010 © 2010 Société statistique du Canada [source]

SCALE MIXTURES DISTRIBUTIONS IN STATISTICAL MODELLING

AUSTRALIAN & NEW ZEALAND JOURNAL OF STATISTICS, Issue 2 2008
S.T. Boris Choy
Summary This paper presents two types of symmetric scale mixture probability distributions which include the normal, Student t, Pearson Type VII, variance gamma, exponential power, uniform power and generalized t (GT) distributions. Expressing a symmetric distribution into a scale mixture form enables efficient Bayesian Markov chain Monte Carlo (MCMC) algorithms in the implementation of complicated statistical models. Moreover, the mixing parameters, a by-product of the scale mixture representation, can be used to identify possible outliers. This paper also proposes a uniform scale mixture representation for the GT density, and demonstrates how this density representation alleviates the computational burden of the Gibbs sampler. [source]