Publication detail

Sequential Monte Carlo estimation of transition probabilities in mixture filtering problems

PAPEŽ, M.

Original Title

Sequential Monte Carlo estimation of transition probabilities in mixture filtering problems

English Title

Sequential Monte Carlo estimation of transition probabilities in mixture filtering problems

Type

conference paper

Language

en

Original Abstract

Physical systems switching between various working regimes are often encountered in practical applications. However, transition probabilities, according to which a system switches from the current regime to another one, are commonly designed as a priori known parameters, and their misspecification can degrade the performance of the algorithms filtering (or estimating) latent variables of the system. To overcome the misspecification, the present paper proposes a novel Sequential Monte Carlo procedure for estimating the transition probabilities. More specifically, it extends the concept of Rao-Blackwellization to the Dirichlet distribution, which represents the model of these probabilities. The experiments show that the proposed technique outperforms some of the classical methods in terms of the estimation precision and also the precision stability.

English abstract

Physical systems switching between various working regimes are often encountered in practical applications. However, transition probabilities, according to which a system switches from the current regime to another one, are commonly designed as a priori known parameters, and their misspecification can degrade the performance of the algorithms filtering (or estimating) latent variables of the system. To overcome the misspecification, the present paper proposes a novel Sequential Monte Carlo procedure for estimating the transition probabilities. More specifically, it extends the concept of Rao-Blackwellization to the Dirichlet distribution, which represents the model of these probabilities. The experiments show that the proposed technique outperforms some of the classical methods in terms of the estimation precision and also the precision stability.

Keywords

Sequential Monte Carlo methods, Rao-Blackwellized particle filter, probabilistic mixtures, switching state-space models

Released

04.08.2016

Publisher

International Society of Information Fusion

Location

Heidelberg

ISBN

978-1-5090-2012-6

Book

Proceedings of the 19th International Conference on Information Fusion, FUSION 2016

Pages from

1063

Pages to

1070

Pages count

8

URL

Documents

BibTex


@inproceedings{BUT127524,
  author="Milan {Papež}",
  title="Sequential Monte Carlo estimation of transition probabilities in mixture filtering problems",
  annote="Physical systems switching between various working regimes are often encountered in practical applications. However, transition probabilities, according to which a system switches from the current regime to another one, are commonly designed as a priori known parameters, and their misspecification can degrade the performance of the algorithms filtering (or estimating) latent variables of the system. To overcome the misspecification, the present paper proposes a novel Sequential Monte Carlo procedure for estimating the transition probabilities. More specifically, it extends the concept of Rao-Blackwellization to the Dirichlet distribution, which represents the model of these probabilities. The experiments show that the proposed technique outperforms some of the classical methods in terms of the estimation precision and also the precision stability.",
  address="International Society of Information Fusion",
  booktitle="Proceedings of the 19th International Conference on Information Fusion, FUSION 2016",
  chapter="127524",
  howpublished="online",
  institution="International Society of Information Fusion",
  year="2016",
  month="august",
  pages="1063--1070",
  publisher="International Society of Information Fusion",
  type="conference paper"
}