Inference for Deterministic Simulation Models: The Bayesian Melding Approach

By David Poole, Adrian E. Raftery. 2000

405 citations recorded by the source corpus.

Subfields: Management Science and Operations Research, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/53a5303f86e2bb6e97d6f04a995410575ccd6c02

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Management Science and Operations Research

In Statistics and Probability

Citations in the corpus

Characterising performance of environmental models (2013, background): “Bayesian techniques for parameter estimation and data assimilation can allow for this by treating both the model and the data as priors with different degrees of uncertainty (Poole and Raftery, 2000).”
Fitting Population Dynamics Models to Count and Cull Data Using Sequential Importance Sampling (2000, methodology): “They allowed prior information on “output parameters” and on “input parameters” so the method is not Bayesian in the usual sense and must be adapted to avoid Borel’s paradox (Poole and Raftery 1998; Raftery and Poole 1997; Raftery et al. 1996).”
Issues in spatially explicit statistical land-use/cover change (LUCC) models: Examples from western (2007, background): “In the Bayesian literature, assimilation of spatially or temporally misaligned data is known as melding, which can overcome a host of potential measurement biases, and missing data (Poole and Raftery, 2000; Sneddon, 2000).”
Demographic research with non-representative internet data (2015, methodology): “Thus, conditional on the model for the bias, we can make statistical inference for the quantities of interest using Bayesian techniques like the Bayesian melding (Poole and Raftery, 2000; Ševčíková et al., 2007; Alkema et al., 2008).”
Bayesian calibration of computer models (2001, methodology): “This was criticised by Wolpert (1995) and Schweder and Hjort (1996), and in a follow-up paper, Poole and Raftery (1998) propose an alternative called Bayesian melding.”
A Spatio-Temporal Downscaler for Output From Numerical Models (2010, methodology): “The model is an application of the Bayesian melding method developed by Poole and Raftery (2000) and builds on the work of Cowles et al. (2002) and Cowles and Zimmerman (2003), who used systematic sampling and numerical integration techniques to combine point and areal data.”
Confidence and Likelihood * (2002, methodology): “With more prior distributions than there are free parameters, Poole & Raftery (1998) propose to meld the priors to a joint prior distribution of the same dimensionality as the free parameter.”
Data integration model for air quality: a hierarchical approach to the global estimation of exposures to ambient air pollution (2016, background): “An alternative approach to the calibration used in downscaling is Bayesian melding (Poole and Raftery, 2000) in which both the measurements and the estimates are assumed to arise from an underlying latent process that represents the true level of the pollutant.”
A Bayesian network of eutrophication models for synthesis, prediction, and uncertainty analysis (2004, neutral): “While progress has been made recently in characterising the uncertainty of complex models (Poole and Raftery, 2000; Kennedy and O’Hagan, 2001; Reichert et al., 2002), most commonly used aquatic ecosystem models have not undergone a rigorous uncertainty analysis (Reckhow, 1994c).”
Covariance Tapering for Likelihood-Based Estimation in Large Spatial Data Sets (2008, neutral): “But this is O(r − ξ ) , because k ∈ ( 0 , 1 ) and ξ < 2 .”