Abstract Dynamic Bayesian models are developed for application in nonlinear, non-normal time series and regression problems, providing dynamic extensions of standard generalized linear models. A key feature the analysis is use conjugate prior posterior distributions exponential family parameters. This leads to calculation closed, standard-form predictive forecasting model criticism. The structure depends on evolution underlying state variables, feedback observational information these...

We develop a Bayesian analysis based on two different Jeffreys priors for the Student-t regression model with unknown degrees of freedom. It is typically difficult to estimate number freedom: improper prior distributions may lead posterior distributions, whereas proper dominate analysis. show that either considered provides distribution. Finally, we estimators compare favourably other previously proposed in literature.

SUMMARY An analysis of a time series cross-sectional data is considered under Bayesian perspective. Information modelled in terms prior distributions and stratified parametric linear models developed by Lindley Smith dynamic Harrison Stevens are merged into general framework. This shown to include many proposed econometrics experimental design. Properties the model derived shrinkage estimators reassessed. Evolution, smoothing passage information through levels hierarchy discussed. Inference...

Abstract Dynamic Bayesian models are developed for application in nonlinear, non-normal time series and regression problems, providing dynamic extensions of standard generalized linear models. A key feature the analysis is use conjugate prior posterior distributions exponential family parameters. This leads to calculation closed, standard-form predictive forecasting model criticism. The structure depends on evolution underlying state variables, feedback observational information these...

Neste artigo, desenvolvemos inferência Bayesiana completa para a classe de modelos DSGE (Dynamics Stochastic General Equilibrium). É bem conhecido que qualquer modelo DSGE, após ser log-linearizado, pode expresso como um espaço estados com choques gaussianos na equação estado. Nesse contexto, propomos uma mistura distribuições normal-gama obter distribuição t-Student correlacionada e abordagem inovadora generaliza multivariada. As metodologias são aplicadas dados reais comparadas tradicional.

We consider models for spatio-temporal processes which assume either non-negative values, and often are observed as zero, or discrete values also inflated by zeros. Typically, in the first case, spatial observations obtained at fixed locations (point-referenced data) over a region D; whereas second, D is divided into finite number of regular irregular subregions (areal level), resulting each subregion. Our main idea based on those zeroinflated models, assuming that value location s time t, Y...

Abstract This paper aims to show practitioners how flexible and straightforward the implementation of Bayesian paradigm can be for distributed lag models within dynamic linear model framework. Distributed are importance when it is believed that a covariate at time t , say X causes an impact on mean value response variable, Y . Moreover, effect persists period decays zero as passes by. There in literature many different deal with this kind situation. review some these proposals under fairly...

The paper introduces a new class of models, named dynamic quantile linear which combines models with distribution-free regression producing robust statistical method. Bayesian estimation for the model is performed using an efficient Markov chain Monte Carlo algorithm. also proposes fast sequential procedure suited high-dimensional predictive modeling massive data, where generating process changing over time. proposed evaluated synthetic and well-known time series data. applied to predict...

This article presents a logistic hierarchical model approach for small area prediction of proportions, taking into account both possible spatial and unstructured heterogeneity effects. The posterior distributions the proportion predictors are obtained via Markov Chain Monte Carlo methods. automatically takes extra uncertainty associated with hyperparameters. procedures applied to real data set comparisons made under several settings, including quite general structure plus A selection...

A stochastic volatility in mean (SVM) model using the class of symmetric scale mixtures normal (SMN) distributions is introduced this article. The SMN form a thick-tailed that includes one as special case, providing robust alternative to estimation SVM models absence normality. Bayesian method via Markov-chain Monte Carlo (MCMC) techniques used estimate parameters. deviance information criterion (DIC) and predictive criteria (BPIC) are calculated compare fit distributions. illustrated by...

This paper introduces electricity load curve models for short‐term forecasting purposes. A broad class of multivariate dynamic regression is proposed to model hourly load. Alternative models, special cases our general model, include separate time series regressions each hour and week day. All the developed components that represent trends, seasons at different levels (yearly, weekly, etc.), dummies take into account weekends/holidays other days, dynamics weather effects, discussing necessity...