Simulating conditionally specified models

Kuo, Kun-Lin; Wang, Yuchung J.

doi:10.1016/j.jmva.2018.04.012

Cited by 2 publications

(5 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To transform the Gibbs sampler into the PCG sampler in a valid sampling order, van Dyk and Park (2008) propose three tools such as marginalization, permutation, and trimming. In addition, Kuo and Wang (2018) show the necessary condition of Gibbs sampling based on the nested sequence of conditional distributions in (5). In particular, the nested condition of ; = A 1 ⊆ A 2 ⊆ Á Á Á ⊆ A n is required for Gibbs sampling in an order of y n !…”

Section: Partial Collapsingmentioning

confidence: 99%

“…becomes functionally incompatible, so that a valid sampling order is required and the PCG sampler must be constructed. Based on the nested condition (Kuo & Wang, 2018), a valid sampling order is either x 1 ! x 3 !…”

Section: Illustrationsmentioning

confidence: 99%

“…The resulting partially collapsed Gibbs (PCG) sampler is extended to the MH-within-PCG sampler ( van Dyk & Jiao, 2015), where certain sampling steps of the PCG sampler are replaced with MH steps while maintaining the desired stationary distribution. The valid sampling order for the PCG sampler is closely investigated by Kuo and Wang (2018).…”

Section: Introductionmentioning

confidence: 99%

“…Note that this PCG sampler cannot be reduced to the CG sampler, as in the previous sampler. Based on the nested condition (Kuo & Wang, 2018), there exists only one valid sampling order, that is, x 1 ! x 3 !…”

mentioning

confidence: 99%

See 3 more Smart Citations

Improving the Gibbs sampler

Park

Lee

2021

WIREs Computational Stats

View full text Add to dashboard Cite

The Gibbs sampler is a simple but very powerful algorithm used to simulate from a complex high-dimensional distribution. It is particularly useful in Bayesian analysis when a complex Bayesian model involves a number of model parameters and the conditional posterior distribution of each component given the others can be derived as a standard distribution. In the presence of a strong correlation structure among components, however, the Gibbs sampler can be criticized for its slow convergence. Here we discuss several algorithmic strategies such as blocking, collapsing, and partial collapsing that are available for improving the convergence characteristics of the Gibbs sampler. This article is categorized under: Statistical and Graphical Methods of Data Analysis > Markov Chain Monte Carlo Statistical and Graphical Methods of Data Analysis > Bayesian Methods and Theory Statistical and Graphical Methods of Data Analysis > Sampling

show abstract

Section: Partial Collapsingmentioning

confidence: 99%

Section: Illustrationsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 99%

See 2 more Smart Citations

Improving the Gibbs sampler

Park

Lee

2021

WIREs Computational Stats

View full text Add to dashboard Cite

show abstract

“…Although the causal and predictive models suggest incompatible distributions for the missing counterfactual outcomes, the use of incompatible imputation and analysis models has been well-studied. This approach, often referred to as a fully conditional model specification, incompatible Markov Chain Monte Carlo (MCMC), or multiple imputation with incongenial sources of input, has been shown to often perform better than fully Bayesian data augmentation with complex models (Meng, 1994;Rubin, 2003;Schafer, 2003;Van Buuren et al, 2006;Van Buuren, 2007;Kuo and Wang, 2018).…”

Section: S4 Bayesian Modularizationmentioning

confidence: 99%

Integrated causal-predictive machine learning models for tropical cyclone epidemiology

Nethery¹,

Katz-Christy²,

Kioumourtzoglou³

et al. 2020

Preprint

View full text Add to dashboard Cite

Strategic preparedness has been shown to reduce the adverse health impacts of hurricanes and tropical storms, referred to collectively as tropical cyclones (TCs), but its protective impact could be enhanced by a more comprehensive and rigorous characterization of TC epidemiology. To generate the insights and tools necessary for high-precision TC preparedness, we develop and apply a novel Bayesian machine learning approach that standardizes estimation of historic TC health impacts, discovers common patterns and sources of heterogeneity in those health impacts, and enables identification of communities at highest health risk for future TCs. The model integrates (1) a causal inference component to quantify the immediate health impacts of recent historic TCs at high spatial resolution and (2) a predictive component that captures how TC meteorological features and socioeconomic/demographic characteristics of impacted communities are associated with health impacts. We apply it to a rich data platform containing detailed historic TC exposure information and Medicare claims data. The health outcomes used in our analyses are all-cause mortality and cardiovascular-and respiratory-related hospitalizations. We report a high degree of heterogeneity in the acute health impacts of historic TCs at both the TC level and the community level, with substantial increases in respiratory hospitalizations, on average, during a two-week period surrounding TCs. TC sustained windspeeds are found to be the primary driver of increased mortality and respiratory risk. Our modeling approach has broader utility for predicting the health impacts of many types of extreme climate events.

show abstract

Simulating conditionally specified models

Cited by 2 publications

References 12 publications

Improving the Gibbs sampler

Improving the Gibbs sampler

Integrated causal-predictive machine learning models for tropical cyclone epidemiology

Contact Info

Product

Resources

About