Stochastic Complexity for tree models

Jun'Ichi Takeuchi, Andrew R. Barron

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

We study the problem of data compression, gambling and prediction of strings xn = x1x2...xn in terms of coding regret, where the tree model is assumed as a target class. We apply the minimax Bayes strategy for curved exponential families to this problem and show that it achieves the minimax regret without restriction on the data strings. This is an extension of the minimax result by (Takeuchi et al. 2013) for models of kth order Markov chains and determines the constant term of the Stochastic Complexity for the tree model.

Original languageEnglish
Title of host publication2014 IEEE Information Theory Workshop, ITW 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages222-226
Number of pages5
ISBN (Electronic)9781479959990
DOIs
Publication statusPublished - Dec 1 2014
Event2014 IEEE Information Theory Workshop, ITW 2014 - Hobart, Australia
Duration: Nov 2 2014Nov 5 2014

Publication series

Name2014 IEEE Information Theory Workshop, ITW 2014

Other

Other2014 IEEE Information Theory Workshop, ITW 2014
CountryAustralia
CityHobart
Period11/2/1411/5/14

Fingerprint

Data compression
Markov processes

All Science Journal Classification (ASJC) codes

  • Information Systems
  • Computer Networks and Communications

Cite this

Takeuchi, JI., & Barron, A. R. (2014). Stochastic Complexity for tree models. In 2014 IEEE Information Theory Workshop, ITW 2014 (pp. 222-226). [6970825] (2014 IEEE Information Theory Workshop, ITW 2014). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ITW.2014.6970825

Stochastic Complexity for tree models. / Takeuchi, Jun'Ichi; Barron, Andrew R.

2014 IEEE Information Theory Workshop, ITW 2014. Institute of Electrical and Electronics Engineers Inc., 2014. p. 222-226 6970825 (2014 IEEE Information Theory Workshop, ITW 2014).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Takeuchi, JI & Barron, AR 2014, Stochastic Complexity for tree models. in 2014 IEEE Information Theory Workshop, ITW 2014., 6970825, 2014 IEEE Information Theory Workshop, ITW 2014, Institute of Electrical and Electronics Engineers Inc., pp. 222-226, 2014 IEEE Information Theory Workshop, ITW 2014, Hobart, Australia, 11/2/14. https://doi.org/10.1109/ITW.2014.6970825
Takeuchi JI, Barron AR. Stochastic Complexity for tree models. In 2014 IEEE Information Theory Workshop, ITW 2014. Institute of Electrical and Electronics Engineers Inc. 2014. p. 222-226. 6970825. (2014 IEEE Information Theory Workshop, ITW 2014). https://doi.org/10.1109/ITW.2014.6970825
Takeuchi, Jun'Ichi ; Barron, Andrew R. / Stochastic Complexity for tree models. 2014 IEEE Information Theory Workshop, ITW 2014. Institute of Electrical and Electronics Engineers Inc., 2014. pp. 222-226 (2014 IEEE Information Theory Workshop, ITW 2014).
@inproceedings{5fdd353deada48dc8d0d56ba6fac7879,
title = "Stochastic Complexity for tree models",
abstract = "We study the problem of data compression, gambling and prediction of strings xn = x1x2...xn in terms of coding regret, where the tree model is assumed as a target class. We apply the minimax Bayes strategy for curved exponential families to this problem and show that it achieves the minimax regret without restriction on the data strings. This is an extension of the minimax result by (Takeuchi et al. 2013) for models of kth order Markov chains and determines the constant term of the Stochastic Complexity for the tree model.",
author = "Jun'Ichi Takeuchi and Barron, {Andrew R.}",
year = "2014",
month = "12",
day = "1",
doi = "10.1109/ITW.2014.6970825",
language = "English",
series = "2014 IEEE Information Theory Workshop, ITW 2014",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "222--226",
booktitle = "2014 IEEE Information Theory Workshop, ITW 2014",
address = "United States",

}

TY - GEN

T1 - Stochastic Complexity for tree models

AU - Takeuchi, Jun'Ichi

AU - Barron, Andrew R.

PY - 2014/12/1

Y1 - 2014/12/1

N2 - We study the problem of data compression, gambling and prediction of strings xn = x1x2...xn in terms of coding regret, where the tree model is assumed as a target class. We apply the minimax Bayes strategy for curved exponential families to this problem and show that it achieves the minimax regret without restriction on the data strings. This is an extension of the minimax result by (Takeuchi et al. 2013) for models of kth order Markov chains and determines the constant term of the Stochastic Complexity for the tree model.

AB - We study the problem of data compression, gambling and prediction of strings xn = x1x2...xn in terms of coding regret, where the tree model is assumed as a target class. We apply the minimax Bayes strategy for curved exponential families to this problem and show that it achieves the minimax regret without restriction on the data strings. This is an extension of the minimax result by (Takeuchi et al. 2013) for models of kth order Markov chains and determines the constant term of the Stochastic Complexity for the tree model.

UR - http://www.scopus.com/inward/record.url?scp=84929316288&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84929316288&partnerID=8YFLogxK

U2 - 10.1109/ITW.2014.6970825

DO - 10.1109/ITW.2014.6970825

M3 - Conference contribution

AN - SCOPUS:84929316288

T3 - 2014 IEEE Information Theory Workshop, ITW 2014

SP - 222

EP - 226

BT - 2014 IEEE Information Theory Workshop, ITW 2014

PB - Institute of Electrical and Electronics Engineers Inc.

ER -