An efficient forward–reverse expectation-maximization algorithm for statistical inference in stochastic reaction networks

Handle URI:
http://hdl.handle.net/10754/621486
Title:
An efficient forward–reverse expectation-maximization algorithm for statistical inference in stochastic reaction networks
Authors:
Bayer, Christian; Moraes, Alvaro ( 0000-0003-4144-1243 ) ; Tempone, Raul ( 0000-0003-1967-4446 ) ; Vilanova, Pedro ( 0000-0001-6620-6261 )
Abstract:
© 2016 Taylor & Francis Group, LLC. ABSTRACT: In this work, we present an extension of the forward–reverse representation introduced by Bayer and Schoenmakers (Annals of Applied Probability, 24(5):1994–2032, 2014) to the context of stochastic reaction networks (SRNs). We apply this stochastic representation to the computation of efficient approximations of expected values of functionals of SRN bridges, that is, SRNs conditional on their values in the extremes of given time intervals. We then employ this SRN bridge-generation technique to the statistical inference problem of approximating reaction propensities based on discretely observed data. To this end, we introduce a two-phase iterative inference method in which, during phase I, we solve a set of deterministic optimization problems where the SRNs are replaced by their reaction-rate ordinary differential equations approximation; then, during phase II, we apply the Monte Carlo version of the expectation-maximization algorithm to the phase I output. By selecting a set of overdispersed seeds as initial points in phase I, the output of parallel runs from our two-phase method is a cluster of approximate maximum likelihood estimates. Our results are supported by numerical examples.
KAUST Department:
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
Citation:
Bayer C, Moraes A, Tempone R, Vilanova P (2016) An efficient forward–reverse expectation-maximization algorithm for statistical inference in stochastic reaction networks. Stochastic Analysis and Applications 34: 193–231. Available: http://dx.doi.org/10.1080/07362994.2015.1116396.
Publisher:
Informa UK Limited
Journal:
Stochastic Analysis and Applications
Issue Date:
20-Feb-2016
DOI:
10.1080/07362994.2015.1116396
Type:
Article
ISSN:
0736-2994; 1532-9356
Sponsors:
The work described in this paper was supported by King Abdullah University of Science and Technology (KAUST). A. Moraes, R. Tempone and P. Vilanova are members of the KAUST SRI Center for Uncertainty Quantification at the Computer, Electrical and Mathematical Science and Engineering Division, KAUST.
Appears in Collections:
Articles; Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division

Full metadata record

DC FieldValue Language
dc.contributor.authorBayer, Christianen
dc.contributor.authorMoraes, Alvaroen
dc.contributor.authorTempone, Raulen
dc.contributor.authorVilanova, Pedroen
dc.date.accessioned2016-11-03T08:30:32Z-
dc.date.available2016-11-03T08:30:32Z-
dc.date.issued2016-02-20en
dc.identifier.citationBayer C, Moraes A, Tempone R, Vilanova P (2016) An efficient forward–reverse expectation-maximization algorithm for statistical inference in stochastic reaction networks. Stochastic Analysis and Applications 34: 193–231. Available: http://dx.doi.org/10.1080/07362994.2015.1116396.en
dc.identifier.issn0736-2994en
dc.identifier.issn1532-9356en
dc.identifier.doi10.1080/07362994.2015.1116396en
dc.identifier.urihttp://hdl.handle.net/10754/621486-
dc.description.abstract© 2016 Taylor & Francis Group, LLC. ABSTRACT: In this work, we present an extension of the forward–reverse representation introduced by Bayer and Schoenmakers (Annals of Applied Probability, 24(5):1994–2032, 2014) to the context of stochastic reaction networks (SRNs). We apply this stochastic representation to the computation of efficient approximations of expected values of functionals of SRN bridges, that is, SRNs conditional on their values in the extremes of given time intervals. We then employ this SRN bridge-generation technique to the statistical inference problem of approximating reaction propensities based on discretely observed data. To this end, we introduce a two-phase iterative inference method in which, during phase I, we solve a set of deterministic optimization problems where the SRNs are replaced by their reaction-rate ordinary differential equations approximation; then, during phase II, we apply the Monte Carlo version of the expectation-maximization algorithm to the phase I output. By selecting a set of overdispersed seeds as initial points in phase I, the output of parallel runs from our two-phase method is a cluster of approximate maximum likelihood estimates. Our results are supported by numerical examples.en
dc.description.sponsorshipThe work described in this paper was supported by King Abdullah University of Science and Technology (KAUST). A. Moraes, R. Tempone and P. Vilanova are members of the KAUST SRI Center for Uncertainty Quantification at the Computer, Electrical and Mathematical Science and Engineering Division, KAUST.en
dc.publisherInforma UK Limiteden
dc.subjectbridges for continuous-time Markov chainsen
dc.subjectForward–reverse algorithmen
dc.subjectinference for stochastic reaction networksen
dc.subjectMonte Carlo expectation-maximization algorithmen
dc.titleAn efficient forward–reverse expectation-maximization algorithm for statistical inference in stochastic reaction networksen
dc.typeArticleen
dc.contributor.departmentComputer, Electrical and Mathematical Sciences and Engineering (CEMSE) Divisionen
dc.identifier.journalStochastic Analysis and Applicationsen
dc.contributor.institutionWeierstrass Institute for Applied Analysis and Stochastics, Berlin, Germanyen
kaust.authorMoraes, Alvaroen
kaust.authorTempone, Raulen
kaust.authorVilanova, Pedroen
All Items in KAUST are protected by copyright, with all rights reserved, unless otherwise indicated.