Bayesian Computation Methods for Inference in Stochastic Kinetic Models
Joint Authors
Koblents, Eugenia
Mariño, Inés P.
Míguez, J.
Source
Issue
Vol. 2019, Issue 2019 (31 Dec. 2019), pp.1-15, 15 p.
Publisher
Hindawi Publishing Corporation
Publication Date
2019-01-20
Country of Publication
Egypt
No. of Pages
15
Main Subjects
Abstract EN
In this paper we investigate Monte Carlo methods for the approximation of the posterior probability distributions in stochastic kinetic models (SKMs).
SKMs are multivariate Markov jump processes that model the interactions among species in biological systems according to a set of usually unknown parameters.
The tracking of the species populations together with the estimation of the interaction parameters is a Bayesian inference problem for which Markov chain Monte Carlo (MCMC) methods have been a typical computational tool.
Specifically, the particle MCMC (pMCMC) method has been shown to be effective, while computationally demanding method applicable to this problem.
Recently, it has been shown that an alternative approach to Bayesian computation, namely, the class of adaptive importance samplers, may be more efficient than classical MCMC-like schemes, at least for certain applications.
For example, the nonlinear population Monte Carlo (NPMC) algorithm has yielded promising results with a low dimensional SKM (the classical predator-prey model).
In this paper we explore the application of both pMCMC and NPMC to analyze complex autoregulatory feedback networks modelled by SKMs.
We demonstrate numerically how the populations of the relevant species in the network can be tracked and their interaction rates estimated, even in scenarios with partial observations.
NPMC schemes attain an appealing trade-off between accuracy and computational cost that can make them advantageous in many practical applications.
American Psychological Association (APA)
Koblents, Eugenia& Mariño, Inés P.& Míguez, J.. 2019. Bayesian Computation Methods for Inference in Stochastic Kinetic Models. Complexity،Vol. 2019, no. 2019, pp.1-15.
https://search.emarefa.net/detail/BIM-1132628
Modern Language Association (MLA)
Koblents, Eugenia…[et al.]. Bayesian Computation Methods for Inference in Stochastic Kinetic Models. Complexity No. 2019 (2019), pp.1-15.
https://search.emarefa.net/detail/BIM-1132628
American Medical Association (AMA)
Koblents, Eugenia& Mariño, Inés P.& Míguez, J.. Bayesian Computation Methods for Inference in Stochastic Kinetic Models. Complexity. 2019. Vol. 2019, no. 2019, pp.1-15.
https://search.emarefa.net/detail/BIM-1132628
Data Type
Journal Articles
Language
English
Notes
Includes bibliographical references
Record ID
BIM-1132628