Skip to main content

Lund University Publications

LUND UNIVERSITY LIBRARIES

Tuned iterated filtering

Lindström, Erik LU orcid (2013) In Statistics and Probability Letters 83(9). p.2077-2080
Abstract
Iterated filtering is an algorithm for estimating parameters in partially observed Markov process (POMP) models. The real-world performance of the algorithm depends on several tuning parameters. We propose a simple method for optimizing the parameter governing the joint dynamics of the hidden parameter process (called the Sigma matrix). The tuning is implemented using a fixed-lag sequential Monte Carlo expectation maximization (EM) algorithm. We introduce two different versions of the tuning parameter, the approximately estimated Sigma matrix, and a normalized version of the same matrix. Our simulations show that the finite-sample performance for the normalized matrix outperform the standard iterated filter, while the naive version is... (More)
Iterated filtering is an algorithm for estimating parameters in partially observed Markov process (POMP) models. The real-world performance of the algorithm depends on several tuning parameters. We propose a simple method for optimizing the parameter governing the joint dynamics of the hidden parameter process (called the Sigma matrix). The tuning is implemented using a fixed-lag sequential Monte Carlo expectation maximization (EM) algorithm. We introduce two different versions of the tuning parameter, the approximately estimated Sigma matrix, and a normalized version of the same matrix. Our simulations show that the finite-sample performance for the normalized matrix outperform the standard iterated filter, while the naive version is doing more harm than good. (C) 2013 Elsevier B.V. All rights reserved. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Contribution to journal
publication status
published
subject
keywords
Hidden Markov models, Sequential Monte Carlo methods, Maximum likelihood, estimation
in
Statistics and Probability Letters
volume
83
issue
9
pages
2077 - 2080
publisher
Elsevier
external identifiers
  • wos:000322295000022
  • scopus:84879164818
ISSN
0167-7152
DOI
10.1016/j.spl.2013.05.019
language
English
LU publication?
yes
id
ead07296-b2d1-4235-866b-339201f58eb3 (old id 4042666)
date added to LUP
2016-04-01 14:51:28
date last changed
2022-01-28 02:53:23
@article{ead07296-b2d1-4235-866b-339201f58eb3,
  abstract     = {{Iterated filtering is an algorithm for estimating parameters in partially observed Markov process (POMP) models. The real-world performance of the algorithm depends on several tuning parameters. We propose a simple method for optimizing the parameter governing the joint dynamics of the hidden parameter process (called the Sigma matrix). The tuning is implemented using a fixed-lag sequential Monte Carlo expectation maximization (EM) algorithm. We introduce two different versions of the tuning parameter, the approximately estimated Sigma matrix, and a normalized version of the same matrix. Our simulations show that the finite-sample performance for the normalized matrix outperform the standard iterated filter, while the naive version is doing more harm than good. (C) 2013 Elsevier B.V. All rights reserved.}},
  author       = {{Lindström, Erik}},
  issn         = {{0167-7152}},
  keywords     = {{Hidden Markov models; Sequential Monte Carlo methods; Maximum likelihood; estimation}},
  language     = {{eng}},
  number       = {{9}},
  pages        = {{2077--2080}},
  publisher    = {{Elsevier}},
  series       = {{Statistics and Probability Letters}},
  title        = {{Tuned iterated filtering}},
  url          = {{http://dx.doi.org/10.1016/j.spl.2013.05.019}},
  doi          = {{10.1016/j.spl.2013.05.019}},
  volume       = {{83}},
  year         = {{2013}},
}