Gaussian Process Regression (GPR) is a powerful tool for nonparametric regression, but its fully Bayesian application in high-dimensional settings hindered by two primary challenges: the computational burden (exacerbated inference) and difficulty of variable selection. This paper introduces novel methodology that combines hierarchical global-local shrinkage priors with normalizing flows to address these challenges. The triple gamma prior offers principled framework inducing sparsity GPR,...

Time-varying parameter (TVP) models are very flexible in capturing gradual changes the effect of explanatory variables on outcome variable. However, particular when number is large, there a known risk overfitting and poor predictive performance, since some constant over time. We propose new prior for variance shrinkage TVP models, called triple gamma. The gamma encompasses priors that have been suggested previously, such as Bayesian Lasso, double Horseshoe prior. present desirable properties...

Time-varying parameter (TVP) models are widely used in time series analysis to flexibly deal with processes which gradually change over time. However, the risk of overfitting TVP is well known. This issue can be dealt using appropriate global-local shrinkage priors, pull time-varying parameters towards static ones. In this paper, we introduce R package shrinkTVP (Knaus, Bitto-Nemling, Cadonna, and FrühwirthSchnatter 2021), provides a fully Bayesian implementation priors for models, taking...

Countries' agricultural systems have an important impact on biodiversity, for example bird populations. Here, we estimate such impacts by exploiting a natural experiment in the middle of Europe, where there is naturally homogenous area that divided into three countries: Switzerland, Germany, and France. These countries markedly different policies. Using methodologically unified unusually rich dataset available across these borders, both 2010s 1990s, analyze (a) whether clear pattern...

Male-biased operational sex ratios are very common in sexually mature dragonflies.These may be due to differential survival or differences time spent at the breeding site by sexes.Because most studies carried out site, these two processes can measured as rates recapture using modern capturemark-recapture methods.We marked 66 female and 233 male Coenagrion puella, 137 347 Ischnura elegans during three capture periods spread over 18 days.Each an animal was recaptured it remarked so that...

Many current approaches to shrinkage within the time-varying parameter framework assume that each state is equipped with only one innovation variance for all time points. Sparsity then induced by shrinking this towards zero. We argue not sufficient if states display large jumps or structural changes, something which often case in series analysis. To remedy this, we propose dynamic triple gamma prior, a stochastic process has well-known marginal form, while still allowing autocorrelation....

Time-varying parameter (TVP) models are widely used in time series analysis to flexibly deal with processes which gradually change over time. However, the risk of overfitting TVP is well known. This issue can be dealt using appropriate global-local shrinkage priors, pull time-varying parameters towards static ones. In this paper, we introduce R package shrinkTVP (Knaus, Bitto-Nemling, Cadonna, and Fr\"uhwirth-Schnatter 2019), provides a fully Bayesian implementation priors for models, taking...

Time-varying parameter (TVP) models are very flexible in capturing gradual changes the effect of a predictor on outcome variable. However, particular when number predictors is large, there known risk overfitting and poor predictive performance, since some constant over time. We propose prior for variance shrinkage TVP models, called triple gamma. The gamma encompasses priors that have been suggested previously, such as Bayesian lasso, double Horseshoe prior. present desirable properties its...

Dynamic survival models are a flexible tool for overcoming limitations of popular methods in the field analysis. While this flexibility allows them to uncover more intricate relationships between covariates and time-to-event, it also has running risk overfitting. This paper proposes solution issue based on state art global-local shrinkage priors shows that they able effectively regularize amount time-variation observed parameters. Further, novel approach accounting unobserved heterogeneity...

In this chapter, we review variance selection for time-varying parameter (TVP) models univariate and multivariate time series within a Bayesian framework. We show how both continuous as well discrete spike-and-slab shrinkage priors can be transferred from variable regression to TVP by using non-centered parametrization. discuss efficient MCMC estimation provide an application US inflation modeling.