Scalable Bayesian inference for self-excitatory stochastic processes applied to big American gunfire data

DOI10.1007/S11222-020-09980-4zbMATH Open1461.62018arXiv2005.10123OpenAlexW3120649110WikidataQ125940879 ScholiaQ125940879MaRDI QIDQ2029080FDOQ2029080

Authors: Andrew J. Holbrook, Charles E. Loeffler, Seth R. Flaxman, Marc A. Suchard 

Publication date: 3 June 2021

Published in: Statistics and Computing (Search for Journal in Brave)

Abstract: The Hawkes process and its extensions effectively model self-excitatory phenomena including earthquakes, viral pandemics, financial transactions, neural spike trains and the spread of memes through social networks. The usefulness of these stochastic process models within a host of economic sectors and scientific disciplines is undercut by the processes' computational burden: complexity of likelihood evaluations grows quadratically in the number of observations for both the temporal and spatiotemporal Hawkes processes. We show that, with care, one may parallelize these calculations using both central and graphics processing unit implementations to achieve over 100-fold speedups over single-core processing. Using a simple adaptive Metropolis-Hastings scheme, we apply our high-performance computing framework to a Bayesian analysis of big gunshot data generated in Washington D.C. between the years of 2006 and 2019, thereby extending a past analysis of the same data from under 10,000 to over 85,000 observations. To encourage wide-spread use, we provide hpHawkes, an open-source R package, and discuss high-level implementation and program design for leveraging aspects of computational hardware that become necessary in a big data setting.

Full work available at URL: https://arxiv.org/abs/2005.10123

Recommendations

• Bayesian mitigation of spatial coarsening for a Hawkes model applied to gunfire, wildfire and viral contagion
• Bayesian inference for Hawkes processes
• An ephemerally self-exciting point process
• Fifty years later: new directions in Hawkes processes
• Mean-field inference of Hawkes point processes

zbMATH Keywords

GPUmassive parallelizationSIMDspatiotemporal Hawkes process

Mathematics Subject Classification ID

Computational methods for problems pertaining to statistics (62-08)

Cites Work

• Power-law models for infectious disease spread
• ggplot2. Elegant graphics for data analysis. With contributions by Carson Sievert
• An Introduction to the Theory of Point Processes
• Spectra of some self-exciting and mutually exciting point processes
• Title not available (Why is that?)
• An adaptive Metropolis algorithm
• Monte Carlo sampling methods using Markov chains and their applications
• Multivariate Hawkes processes: an application to financial data
• Equation of state calculations by fast computing machines
• Coupling and Ergodicity of Adaptive Markov Chain Monte Carlo Algorithms
• A review of self-exciting spatio-temporal point processes and their applications
• An Introduction to the Theory of Point Processes
• Bayesian inference for Hawkes processes
• Title not available (Why is that?)
• Graphics processing units and high-dimensional optimization
• GPU accelerated MCMC for modeling terrorist activity

Cited In (6)

• Approximation of Bayesian Hawkes process with \texttt{inlabru}
• Bayesian mitigation of spatial coarsening for a Hawkes model applied to gunfire, wildfire and viral contagion
• A Quantum Parallel Markov Chain Monte Carlo
• Fifty years later: new directions in Hawkes processes
• hpHawkes
• Improvements on scalable stochastic Bayesian inference methods for multivariate Hawkes process

Uses Software

• R
• CODA
• Rcpp
• ggplot2
• Intel TBB
• RcppParallel
• PyHawkes
• RcppNT2
• MassiveMDS

This page was built for publication: Scalable Bayesian inference for self-excitatory stochastic processes applied to big American gunfire data

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2029080)