Bayesian design of experiments for intractable likelihood models using coupled auxiliary models and multivariate emulation

DOI10.1214/19-BA1144zbMATH Open1437.62303OpenAlexW2963983664WikidataQ128410751 ScholiaQ128410751MaRDI QIDQ2297233FDOQ2297233

Authors: Antony M. Overstall, J. M. McGree 

Publication date: 18 February 2020

Published in: Bayesian Analysis (Search for Journal in Brave)

Abstract: A Bayesian design is given by maximising an expected utility over a design space. The utility is chosen to represent the aim of the experiment and its expectation is taken with respect to all unknowns: responses, parameters and/or models. Although straightforward in principle, there are several challenges to finding Bayesian designs in practice. Firstly, the utility and expected utility are rarely available in closed form and require approximation. Secondly, the design space can be of high-dimensionality. In the case of intractable likelihood models, these problems are compounded by the fact that the likelihood function, whose evaluation is required to approximate the expected utility, is not available in closed form. A strategy is proposed to find Bayesian designs for intractable likelihood models. It relies on the development of an automatic, auxiliary modelling approach, using multivariate Gaussian process emulators, to approximate the likelihood function. This is then combined with a copula-based approach to approximate the marginal likelihood (a quantity commonly required to evaluate many utility functions). These approximations are demonstrated on examples of stochastic process models involving experimental aims of both parameter estimation and model comparison.

Full work available at URL: https://arxiv.org/abs/1803.07018

Recommendations

• Bayesian experimental design for models with intractable likelihoods
• Optimal Bayesian experimental design for models with intractable likelihoods using indirect inference applied to biological process models
• Optimal Bayesian design for discriminating between models with intractable likelihoods in epidemiology
• Bayesian sequential design for copula models
• An approach for finding fully Bayesian optimal designs using normal-based approximations to loss functions

zbMATH Keywords

parameter estimationapproximate Bayesian computationGaussian processindirect inferencemodel comparisonapproximate coordinate exchange

Mathematics Subject Classification ID

Bayesian inference (62F15) Applications of statistics to biology and medical sciences; meta analysis (62P10) Gaussian processes (60G15) Characterization and structure theory for multivariate probability distributions; copulas (62H05) Optimal statistical designs (62K05)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• A Review of Modern Computational Algorithms for Bayesian Optimal Design
• An approach for finding fully Bayesian optimal designs using normal-based approximations to loss functions
• Bayes linear analysis for Bayesian optimal experimental design
• Bayesian data analysis.
• Bayesian design of experiments for intractable likelihood models using coupled auxiliary models and multivariate emulation
• Bayesian emulation of complex multi-output and dynamic computer models
• Bayesian experimental design for models with intractable likelihoods
• Bayesian experimental design: A review
• Bayesian indirect inference using a parametric auxiliary model
• Bayesian-Optimal Design via Interacting Particle Systems
• Computational enhancements to Bayesian design of experiments using Gaussian processes
• Efficient construction of Bayes optimal designs for stochastic process models
• Encompassing and indirect inference
• Estimating Functions in Indirect Inference
• Fast estimation of expected information gains for Bayesian experimental designs based on Laplace approximations
• On a Measure of the Information Provided by an Experiment
• On the efficient determination of optimal Bayesian experimental designs using ABC: a case study in optimal observation of epidemics
• Optimal Bayesian Design by Inhomogeneous Markov Chain Simulation
• Optimal Bayesian design for discriminating between models with intractable likelihoods in epidemiology
• Optimal Bayesian experimental design for models with intractable likelihoods using indirect inference applied to biological process models
• Optimal Design via Curve Fitting of Monte Carlo Experiments
• Optimal design of experimental epidemics
• Optimal design of measurements on queueing systems
• Optimum experimental designs, with SAS
• Pair copula constructions for multivariate discrete data
• Robust parameter estimation techniques for stochastic within-host macroparasite models
• Stochastic modeling of aphid population growth with nonlinear, power-law dynamics
• The Coordinate-Exchange Algorithm for Constructing Exact Optimal Experimental Designs
• The design and analysis of computer experiments.
• The t Copula and Related Copulas
• Towards Bayesian experimental design for nonlinear models that require a large number of sampling times

Cited In (7)

• Optimal Bayesian experimental design for models with intractable likelihoods using indirect inference applied to biological process models
• Optimal Bayesian design for discriminating between models with intractable likelihoods in epidemiology
• Bayesian design of experiments for intractable likelihood models using coupled auxiliary models and multivariate emulation
• Parallel Gaussian process surrogate Bayesian inference with noisy likelihood evaluations
• Bayesian experimental design for models with intractable likelihoods
• Modern Bayesian experimental design
• Optimal Bayesian design for model discrimination via classification

Uses Software

• BayesDA
• acebayes

This page was built for publication: Bayesian design of experiments for intractable likelihood models using coupled auxiliary models and multivariate emulation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2297233)