Home > Research > Publications & Outputs > Direct statistical inference for finite Markov ...

Electronic data

  • MatExp

    Accepted author manuscript, 405 KB, PDF document

    Available under license: CC BY: Creative Commons Attribution 4.0 International License

Links

Text available via DOI:

View graph of relations

Direct statistical inference for finite Markov jump processes via the matrix exponential

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Published

Standard

Direct statistical inference for finite Markov jump processes via the matrix exponential. / Sherlock, Chris.

In: Computational Statistics, Vol. 36, No. 4, 31.12.2021, p. 2863-2887.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

APA

Vancouver

Author

Sherlock, Chris. / Direct statistical inference for finite Markov jump processes via the matrix exponential. In: Computational Statistics. 2021 ; Vol. 36, No. 4. pp. 2863-2887.

Bibtex

@article{8e34ae6125d14696863ff1524219ea54,
title = "Direct statistical inference for finite Markov jump processes via the matrix exponential",
abstract = "Given noisy, partial observations of a time-homogeneous, finite-statespace Markov chain, conceptually simple, direct statistical inference is available, in theory, via its rate matrix, or infinitesimal generator, Q, since exp (Qt) is the transition matrix over time t. However, perhaps because of inadequate tools for matrix exponentiation in programming languages commonly used amongst statisticians or a belief that the necessary calculations are prohibitively expensive, statistical inference for continuous-time Markov chains with a large but finite state space is typically conducted via particle MCMC or other relatively complex inference schemes. When, as in many applications Q arises from a reaction network, it is usually sparse. We describe variations on known algorithms which allow fast, robust and accurate evaluation of the product of a non-negative vector with the exponential of a large, sparse rate matrix. Our implementation uses relatively recently developed, efficient, linear algebra tools that take advantage of such sparsity. We demonstrate the straightforward statistical application of the key algorithm on a model for the mixing of two alleles in a population and on the Susceptible-Infectious-Removed epidemic model.",
keywords = "Markov jump process, Likelihood inference, Bayesian inference, Matrix exponential",
author = "Chris Sherlock",
year = "2021",
month = dec,
day = "31",
doi = "10.1007/s00180-021-01102-6",
language = "English",
volume = "36",
pages = "2863--2887",
journal = "Computational Statistics",
issn = "0943-4062",
publisher = "Springer Verlag",
number = "4",

}

RIS

TY - JOUR

T1 - Direct statistical inference for finite Markov jump processes via the matrix exponential

AU - Sherlock, Chris

PY - 2021/12/31

Y1 - 2021/12/31

N2 - Given noisy, partial observations of a time-homogeneous, finite-statespace Markov chain, conceptually simple, direct statistical inference is available, in theory, via its rate matrix, or infinitesimal generator, Q, since exp (Qt) is the transition matrix over time t. However, perhaps because of inadequate tools for matrix exponentiation in programming languages commonly used amongst statisticians or a belief that the necessary calculations are prohibitively expensive, statistical inference for continuous-time Markov chains with a large but finite state space is typically conducted via particle MCMC or other relatively complex inference schemes. When, as in many applications Q arises from a reaction network, it is usually sparse. We describe variations on known algorithms which allow fast, robust and accurate evaluation of the product of a non-negative vector with the exponential of a large, sparse rate matrix. Our implementation uses relatively recently developed, efficient, linear algebra tools that take advantage of such sparsity. We demonstrate the straightforward statistical application of the key algorithm on a model for the mixing of two alleles in a population and on the Susceptible-Infectious-Removed epidemic model.

AB - Given noisy, partial observations of a time-homogeneous, finite-statespace Markov chain, conceptually simple, direct statistical inference is available, in theory, via its rate matrix, or infinitesimal generator, Q, since exp (Qt) is the transition matrix over time t. However, perhaps because of inadequate tools for matrix exponentiation in programming languages commonly used amongst statisticians or a belief that the necessary calculations are prohibitively expensive, statistical inference for continuous-time Markov chains with a large but finite state space is typically conducted via particle MCMC or other relatively complex inference schemes. When, as in many applications Q arises from a reaction network, it is usually sparse. We describe variations on known algorithms which allow fast, robust and accurate evaluation of the product of a non-negative vector with the exponential of a large, sparse rate matrix. Our implementation uses relatively recently developed, efficient, linear algebra tools that take advantage of such sparsity. We demonstrate the straightforward statistical application of the key algorithm on a model for the mixing of two alleles in a population and on the Susceptible-Infectious-Removed epidemic model.

KW - Markov jump process

KW - Likelihood inference

KW - Bayesian inference

KW - Matrix exponential

U2 - 10.1007/s00180-021-01102-6

DO - 10.1007/s00180-021-01102-6

M3 - Journal article

VL - 36

SP - 2863

EP - 2887

JO - Computational Statistics

JF - Computational Statistics

SN - 0943-4062

IS - 4

ER -