Universal models for the exponential distribution

Daniel F. Schmidt; Enes Makalic

doi:10.1109/TIT.2009.2018331

Universal models for the exponential distribution

Daniel F. Schmidt, Enes Makalic

Research output: Contribution to journal › Article › Research › peer-review

18 Citations (Scopus)

Abstract

This paper considers the problem of constructing information theoretic universal models for data distributed according to the exponential distribution. The universal models examined include the sequential Normalized Maximum Likelihood (SNML) code, conditional normalized maximum likelihood (CNML) code, the minimum message length (MML) code, and the Bayes mixture code (BMC). The CNML code yields a codelength identical to the Bayesian mixture code, and within O(1) of the MML codelength, with suitable data driven priors.

Original language	English
Pages (from-to)	3087-3090
Number of pages	4
Journal	IEEE Transactions on Information Theory
Volume	55
Issue number	7
DOIs	https://doi.org/10.1109/TIT.2009.2018331
Publication status	Published - 2009
Externally published	Yes

Keywords

Minimum description length (MDL)
Minimum message length (MML)
Universal models

Access to Document

10.1109/TIT.2009.2018331

Cite this

@article{db6f9bfc4954485da3b68e41a72b5efc,

title = "Universal models for the exponential distribution",

abstract = "This paper considers the problem of constructing information theoretic universal models for data distributed according to the exponential distribution. The universal models examined include the sequential Normalized Maximum Likelihood (SNML) code, conditional normalized maximum likelihood (CNML) code, the minimum message length (MML) code, and the Bayes mixture code (BMC). The CNML code yields a codelength identical to the Bayesian mixture code, and within O(1) of the MML codelength, with suitable data driven priors.",

keywords = "Minimum description length (MDL), Minimum message length (MML), Universal models",

author = "Schmidt, {Daniel F.} and Enes Makalic",

year = "2009",

doi = "10.1109/TIT.2009.2018331",

language = "English",

volume = "55",

pages = "3087--3090",

journal = "IEEE Transactions on Information Theory",

issn = "0018-9448",

publisher = "IEEE, Institute of Electrical and Electronics Engineers",

number = "7",

}

TY - JOUR

T1 - Universal models for the exponential distribution

AU - Schmidt, Daniel F.

AU - Makalic, Enes

PY - 2009

Y1 - 2009

N2 - This paper considers the problem of constructing information theoretic universal models for data distributed according to the exponential distribution. The universal models examined include the sequential Normalized Maximum Likelihood (SNML) code, conditional normalized maximum likelihood (CNML) code, the minimum message length (MML) code, and the Bayes mixture code (BMC). The CNML code yields a codelength identical to the Bayesian mixture code, and within O(1) of the MML codelength, with suitable data driven priors.

AB - This paper considers the problem of constructing information theoretic universal models for data distributed according to the exponential distribution. The universal models examined include the sequential Normalized Maximum Likelihood (SNML) code, conditional normalized maximum likelihood (CNML) code, the minimum message length (MML) code, and the Bayes mixture code (BMC). The CNML code yields a codelength identical to the Bayesian mixture code, and within O(1) of the MML codelength, with suitable data driven priors.

KW - Minimum description length (MDL)

KW - Minimum message length (MML)

KW - Universal models

UR - http://www.scopus.com/inward/record.url?scp=67650151803&partnerID=8YFLogxK

U2 - 10.1109/TIT.2009.2018331

DO - 10.1109/TIT.2009.2018331

M3 - Article

AN - SCOPUS:67650151803

SN - 0018-9448

VL - 55

SP - 3087

EP - 3090

JO - IEEE Transactions on Information Theory

JF - IEEE Transactions on Information Theory

IS - 7

ER -

Universal models for the exponential distribution

Abstract

Keywords

Access to Document

Other files and links

Cite this