Assessing and Visualizing Simultaneous Simulation Error

Robertson, Nathan; Flegal, James M.; Vats, Dootika; Jones, Galin L.

doi:10.1080/10618600.2020.1824871

Cited by 12 publications

(5 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Often there are many expectations and marginal quantiles of interest, but in the interest of a concise presentation, we will focus on univariate expectations; the general case is more complicated but can be handled analogously (Robertson et al, 2021;Vats et al, 2019).…”

Section: Learning About Fmentioning

confidence: 99%

Rank-Normalization, Folding, and Localization: An Improved Rˆ for Assessing Convergence of MCMC (with Discussion)

et al. 2021

View full text Add to dashboard Cite

Markov chain Monte Carlo is a key computational tool in Bayesian statistics, but it can be challenging to monitor the convergence of an iterative stochastic algorithm. In this paper we show that the convergence diagnostic R of Gelman and Rubin (1992) has serious flaws. Traditional R will fail to correctly diagnose convergence failures when the chain has a heavy tail or when the variance varies across the chains. In this paper we propose an alternative rank-based diagnostic that fixes these problems. We also introduce a collection of quantile-based local efficiency measures, along with a practical approach for computing Monte Carlo error estimates for quantiles. We suggest that common trace plots should be replaced with rank plots from multiple chains. Finally, we give recommendations for how these methods should be used in practice. * We thank Ben Bales, Ian Langmore, the editor, and anonymous reviewers for useful comments.

show abstract

Section: Learning About Fmentioning

confidence: 99%

Rank-Normalization, Folding, and Localization: An Improved Rˆ for Assessing Convergence of MCMC (with Discussion)

et al. 2021

View full text Add to dashboard Cite

show abstract

“…At present, the sensor simulation mode is divided into three ways: physical signal simulation, original data simulation and target list simulation. Physical signal simulation simulates the transmission signal of the sensor signal through professional equipment, such as millimeter wave sensor, and physical signal simulation is realized by simulating the radar electromagnetic echo [6][7][8][9]. The original signal simulation mode refers to the untreated raw data of the sensor input through professional equipment.…”

Section: Research On Test Error In Adas Functional Simulationmentioning

confidence: 99%

Research on hardware in the loop simulation error based on driver assistance function testing

Yu,

Li,

Zheng

et al. 2023

Eighth International Conference on Electromechanical Control Technology and Transportation (ICECTT 2023)

View full text Add to dashboard Cite

“…Convergence analyses of general state space Metropolis–Hastings algorithms have traditionally focused on studying their convergence rates in total variation distances [31, 44, 48]. These convergence rates have received significant attention, at least in part, because they provide a key sufficient condition for the existence of central limit theorems [22] and the validity of methods for assessing the reliability of the simulation effort [43, 49]. However, convergence analyses of Metropolis–Hastings Markov chains typically result in qualitative convergence rates [15, 18, 21, 29, 42].…”

Section: Introductionmentioning

confidence: 99%

Exact convergence analysis for metropolis–hastings independence samplers in Wasserstein distances

Brown

Jones

2023

J. Appl. Probab.

Self Cite

View full text Add to dashboard Cite

Under mild assumptions, we show that the exact convergence rate in total variation is also exact in weaker Wasserstein distances for the Metropolis–Hastings independence sampler. We develop a new upper and lower bound on the worst-case Wasserstein distance when initialized from points. For an arbitrary point initialization, we show that the convergence rate is the same and matches the convergence rate in total variation. We derive exact convergence expressions for more general Wasserstein distances when initialization is at a specific point. Using optimization, we construct a novel centered independent proposal to develop exact convergence rates in Bayesian quantile regression and many generalized linear model settings. We show that the exact convergence rate can be upper bounded in Bayesian binary response regression (e.g. logistic and probit) when the sample size and dimension grow together.

show abstract

Assessing and Visualizing Simultaneous Simulation Error

Cited by 12 publications

References 55 publications

Rank-Normalization, Folding, and Localization: An Improved Rˆ for Assessing Convergence of MCMC (with Discussion)

Rank-Normalization, Folding, and Localization: An Improved Rˆ for Assessing Convergence of MCMC (with Discussion)

Research on hardware in the loop simulation error based on driver assistance function testing

Exact convergence analysis for metropolis–hastings independence samplers in Wasserstein distances

Contact Info

Product

Resources

About