Likelihood inflating sampling algorithm

DOI10.1002/CJS.11343MaRDI QIDQ4960912zbMATH OpenOpenAlexFDO

Authors Reihaneh Entezari, Radu V. Craiu, Jeffrey S. Rosenthal

Publication date 24 April 2020

Published in The Canadian Journal of Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1605.02113

zbMATH Keywords

Bayesian inferenceMarkov chain Monte Carlo (MCMC)big dataconsensus Monte CarloBayesian additive regression trees (BART)

Mathematics Subject Classification ID

Bayesian inference (62F15) Statistical aspects of big data and data science (62R07) Inference from stochastic processes and prediction (62M20)

Abstract: Markov Chain Monte Carlo (MCMC) sampling from a posterior distribution corresponding to a massive data set can be computationally prohibitive since producing one sample requires a number of operations that is linear in the data size. In this paper, we introduce a new communication-free parallel method, the Likelihood Inflating Sampling Algorithm (LISA), that significantly reduces computational costs by randomly splitting the dataset into smaller subsets and running MCMC methods independently in parallel on each subset using different processors. Each processor will be used to run an MCMC chain that samples sub-posterior distributions which are defined using an "inflated" likelihood function. We develop a strategy for combining the draws from different sub-posteriors to study the full posterior of the Bayesian Additive Regression Trees (BART) model. The performance of the method is tested using both simulated and real data.

Recommendations

• Double-parallel Monte Carlo for Bayesian analysis of big data
• Speeding Up MCMC by Efficient Data Subsampling
• Embarrassingly parallel sequential Markov-chain Monte Carlo for large sets of time series
• Most likely optimal subsampled Markov chain Monte Carlo
• Parallel multivariate slice sampling

Cited in

(9)

• Bayesian additive regression trees and the general BART model
• Double-parallel Monte Carlo for Bayesian analysis of big data
• Consensus Monte Carlo for Random Subsets Using Shared Anchors
• Automated scalable Bayesian inference via Hilbert coresets
• Method G: Uncertainty Quantification for Distributed Data Problems Using Generalized Fiducial Inference
• SwISS: a scalable Markov chain Monte Carlo divide-and-conquer strategy
• Finding our way in the dark: approximate MCMC for approximate Bayesian methods
• Reflections on Bayesian inference and Markov chain Monte Carlo
• Adaptive Component-Wise Multiple-Try Metropolis Sampling

This page was built for publication: Likelihood inflating sampling algorithm

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4960912)