Mosquito-borne diseases' impact on human health is among the most prominent of all communicable diseases. With limited pool tools to contrast these diseases, public focus remains preventing mosquito-human contacts. Applying a hierarchical spatio-temporal Bayesian model West Nile virus (WNV) surveillance data from Greece, we aimed investigate climatic and environmental factors Culex mosquitoes' population. Our analysis confirmed as major drivers WNV-transmitting-Culex mosquitoes population...

With a case-fatality-risk ranging from 3.0 to >20.0% and life-long sequelae, West Nile neuroinvasive disease (WNND) is the most dangerous outcome of virus (WNV) infection in humans. As no specific prophylaxis nor therapy available for these infections, focus on preventive strategies. We aimed find variables associated with WNND diagnosis, hospitalisation or death, identify high-risk sub-groups population, whom concentrate strategies.We used data The European Surveillance System-TESSy,...

Summary Modern methods for Bayesian regression beyond the Gaussian response setting are often computationally impractical or inaccurate in high dimensions. In fact, as discussed recent literature, bypassing such a trade-off is still an open problem even routine binary models, and there limited theory on quality of variational approximations high-dimensional settings. To address this gap, we study approximation accuracy routinely used mean-field Bayes solutions probit with priors, obtaining...

We extend a previously developed epidemiological model for West Nile virus (WNV) infection in humans Greece, employing laboratory-confirmed WNV cases and mosquito-specific characteristics of transmission, such as host selection temperature-dependent transmission the virus. Host was defined by bird human selection, latter accounting only fraction that develop symptoms after is acquired. To role temperature on we considered five intervals (≤ 19.25 °C; > < 21.75 ≥ 24.25 26.75 °C). The capacity...

Abstract Non-Gaussian state-space models arise in several applications, and within this framework the binary time series setting provides a relevant example. However, unlike for Gaussian — where filtering, predictive smoothing distributions are available closed form require approximations or sequential Monte Carlo strategies inference prediction. This is due to apparent absence of conjugacy between states likelihood induced by observation equation data. In article we prove that dynamic...

Multinomial probit models are routinely-implemented representations for learning how the class probabilities of categorical response data change with p observed predictors. Although several frequentist methods have been developed estimation, inference and classification within such a models, Bayesian is still lagging behind. This due to apparent absence tractable conjugate priors, that may facilitate posterior on multinomial coefficients. Such an issue has motivated increasing efforts toward...

We congratulate the authors for their methodological and theoretical contribution to statistical literature on networks.A natural extension of proposed PAPER model is included, with K communities growing simultaneously where new nodes are either assigned an existing community or elected as a root.The employed assignment rule Pólya-urn type, which leads logarithmic growth number (Korwar Hollander, 1972) known coincide predictive scheme exchangeable sequences associated Dirichlet process.The...

Generalized linear models (GLMs) arguably represent the standard approach for statistical regression beyond Gaussian likelihood scenario. When Bayesian formulations are employed, general absence of a tractable posterior distribution has motivated development deterministic approximations, which generally more scalable than sampling techniques. Among them, expectation propagation (EP) showed extreme accuracy, usually higher many variational Bayes solutions. However, computational cost EP posed...

Binary regression models represent a popular model-based approach for binary classification. In the Bayesian framework, computational challenges in form of posterior distribution motivate still-ongoing fruitful research. Here, we focus on computation predictive probabilities probit via expectation propagation (EP). Leveraging more general results recent literature, show that such admit closed-form expression. Improvements over state-of-the-art approaches are shown simulation study.

Bayesian binary regression is a prosperous area of research due to the computational challenges encountered by currently available methods either for high-dimensional settings or large datasets, both. In present work, we focus on expectation propagation (EP) approximation posterior distribution in probit under multivariate Gaussian prior distribution. Adapting more general derivations Anceschi et al. (2023), show how leverage results extended skew-normal derive an efficient implementation EP...

The smoothing distribution of dynamic probit models with Gaussian state dynamics was recently proved to belong the unified skew-normal family. Although this is computationally tractable in small-to-moderate settings, it may become impractical higher dimensions. In work, adapting a recent more general class expectation propagation (EP) algorithms, we derive an efficient EP routine perform inference for such distribution. We show that proposed approximation leads accuracy gains over available...

Modern methods for Bayesian regression beyond the Gaussian response setting are often computationally impractical or inaccurate in high dimensions. In fact, as discussed recent literature, bypassing such a trade-off is still an open problem even routine binary models, and there limited theory on quality of variational approximations high-dimensional settings. To address this gap, we study approximation accuracy routinely-used mean-field Bayes solutions probit with priors, obtaining novel...

Multinomial probit (mnp) models are fundamental and widely-applied regression for categorical data. Fasano Durante (2022) proved that the class of unified skew-normal distributions is conjugate to several mnp sampling models. This allows develop Monte Carlo samplers accurate variational methods perform Bayesian inference. In this paper, we adapt abovementioned results a popular special case: discrete-choice model under zero mean independent Gaussian priors. obtain simplified expressions...

Recently, Fasano, Rebaudo, Durante and Petrone (2019) provided closed-form expressions for the filtering, predictive smoothing distributions of multivariate dynamic probit models, leveraging on unified skew-normal distribution properties. This allows to develop algorithms draw independent identically distributed samples from such distributions, as well sequential Monte Carlo procedures filtering allowing overcome computational bottlenecks that may arise large sample sizes. In this paper, we...