Repository: Freie Universität Berlin, Math Department

Observation Uncertainty in Reversible Markov Chains

Metzner, Ph. and Weber, M. and Schütte, Ch. (2010) Observation Uncertainty in Reversible Markov Chains. Phys. Rev. E, 82 (3). 031114.

[img] PDF - Accepted Version
Restricted to Registered users only

9MB

Official URL: http://dx.doi.org/10.1103/PhysRevE.82.031114

Abstract

In many applications one is interested in finding a simplified model which captures the essential dynamical behavior of a real life process. If the essential dynamics can be assumed to be (approximately) memoryless then a reasonable choice for a model is a Markov model whose parameters are estimated by means of Bayesian inference from an observed time series. We propose an efficient Monte Carlo Markov Chain framework to assess the uncertainty of the Markov model and related observables. The derived Gibbs sampler allows for sampling distributions of transition matrices subject to reversibility and/or sparsity constraints. The performance of the suggested sampling scheme is demonstrated and discussed for a variety of model examples. The uncertainty analysis of functions of the Markov model under investigation is discussed in application to the identification of conformations of the trialanine molecule via Robust Perron Cluster Analysis (PCCA+).

Item Type:Article
Subjects:Mathematical and Computer Sciences
Divisions:Department of Mathematics and Computer Science > Institute of Mathematics
Department of Mathematics and Computer Science > Institute of Mathematics > BioComputing Group
ID Code:920
Deposited By: BioComp Admin
Deposited On:28 Jul 2010 14:28
Last Modified:03 Mar 2017 14:40

Repository Staff Only: item control page