Change-point estimation using new Minimum Message Length approximations

Leigh J Fitzgibbon; David L Dowe; Lloyd Allison

doi:10.1007/3-540-45683-X_28

Change-point estimation using new Minimum Message Length approximations

Leigh J Fitzgibbon, David L Dowe, Lloyd Allison

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

14 Citations (Scopus)

Abstract

This paper investigates the coding of change-points in the information-theoretic Minimum Message Length (MML) framework. Change-point coding regions affect model selection and parameter estimation in problems such as time series segmentation and decision trees. The Minimum Message Length (MML) and Minimum Description Length (MDL78) approaches to change-point problems have been shown to perform well by several authors. In this paper we compare some published MML and MDL78 methods and introduce some new MML approximations called ‘MMLDc’ and ‘MMLDF’. These new approximations are empirically compared with Strict MML (SMML), Fairly Strict MML (FSMML), MML68, the Minimum Expected Kullback-Leibler Distance (MEKLD) loss function and MDL78 on a tractable binomial change-point problem.

Original language	English
Title of host publication	PRICAI 2002: Trends in Artificial Intelligence
Subtitle of host publication	7th Pacific Rim International Conference on Artificial Intelligence Tokyo, Japan, August 18-22, 2002 Proceedings
Editors	Mitsuru Ishizuka, Abdul Sattar
Place of Publication	Berlin Germany
Publisher	Springer
Pages	244-254
Number of pages	11
ISBN (Print)	3540440380
DOIs	https://doi.org/10.1007/3-540-45683-X_28
Publication status	Published - 2002
Event	Pacific Rim International Conference on Artificial Intelligence 2002 - Tokyo, Japan Duration: 18 Aug 2002 → 22 Aug 2002 Conference number: 7th https://link.springer.com/book/10.1007/3-540-45683-X (Proceedings)

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer
Volume	2417
ISSN (Print)	0302-9743

Conference

Conference	Pacific Rim International Conference on Artificial Intelligence 2002
Abbreviated title	PRICAI 2002
Country/Territory	Japan
City	Tokyo
Period	18/08/02 → 22/08/02
Internet address	https://link.springer.com/book/10.1007/3-540-45683-X (Proceedings)

Access to Document

10.1007/3-540-45683-X_28

Cite this

Fitzgibbon, L. J., Dowe, D. L., & Allison, L. (2002). Change-point estimation using new Minimum Message Length approximations. In M. Ishizuka, & A. Sattar (Eds.), PRICAI 2002: Trends in Artificial Intelligence: 7th Pacific Rim International Conference on Artificial Intelligence Tokyo, Japan, August 18-22, 2002 Proceedings (pp. 244-254). (Lecture Notes in Computer Science; Vol. 2417). Springer. https://doi.org/10.1007/3-540-45683-X_28

Fitzgibbon, Leigh J ; Dowe, David L ; Allison, Lloyd. / Change-point estimation using new Minimum Message Length approximations. PRICAI 2002: Trends in Artificial Intelligence: 7th Pacific Rim International Conference on Artificial Intelligence Tokyo, Japan, August 18-22, 2002 Proceedings. editor / Mitsuru Ishizuka ; Abdul Sattar. Berlin Germany : Springer, 2002. pp. 244-254 (Lecture Notes in Computer Science).

@inproceedings{39cafe0ed45f40b08cfb704f58bb4f03,

title = "Change-point estimation using new Minimum Message Length approximations",

abstract = "This paper investigates the coding of change-points in the information-theoretic Minimum Message Length (MML) framework. Change-point coding regions affect model selection and parameter estimation in problems such as time series segmentation and decision trees. The Minimum Message Length (MML) and Minimum Description Length (MDL78) approaches to change-point problems have been shown to perform well by several authors. In this paper we compare some published MML and MDL78 methods and introduce some new MML approximations called {\textquoteleft}MMLDc{\textquoteright} and {\textquoteleft}MMLDF{\textquoteright}. These new approximations are empirically compared with Strict MML (SMML), Fairly Strict MML (FSMML), MML68, the Minimum Expected Kullback-Leibler Distance (MEKLD) loss function and MDL78 on a tractable binomial change-point problem.",

author = "Fitzgibbon, {Leigh J} and Dowe, {David L} and Lloyd Allison",

year = "2002",

doi = "10.1007/3-540-45683-X_28",

language = "English",

isbn = "3540440380",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "244--254",

editor = "Mitsuru Ishizuka and Abdul Sattar",

booktitle = "PRICAI 2002: Trends in Artificial Intelligence",

note = "Pacific Rim International Conference on Artificial Intelligence 2002, PRICAI 2002 ; Conference date: 18-08-2002 Through 22-08-2002",

url = "https://link.springer.com/book/10.1007/3-540-45683-X",

}

Fitzgibbon, LJ, Dowe, DL & Allison, L 2002, Change-point estimation using new Minimum Message Length approximations. in M Ishizuka & A Sattar (eds), PRICAI 2002: Trends in Artificial Intelligence: 7th Pacific Rim International Conference on Artificial Intelligence Tokyo, Japan, August 18-22, 2002 Proceedings. Lecture Notes in Computer Science, vol. 2417, Springer, Berlin Germany, pp. 244-254, Pacific Rim International Conference on Artificial Intelligence 2002, Tokyo, Japan, 18/08/02. https://doi.org/10.1007/3-540-45683-X_28

Change-point estimation using new Minimum Message Length approximations. / Fitzgibbon, Leigh J; Dowe, David L; Allison, Lloyd.

PRICAI 2002: Trends in Artificial Intelligence: 7th Pacific Rim International Conference on Artificial Intelligence Tokyo, Japan, August 18-22, 2002 Proceedings. ed. / Mitsuru Ishizuka; Abdul Sattar. Berlin Germany : Springer, 2002. p. 244-254 (Lecture Notes in Computer Science; Vol. 2417).

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

TY - GEN

T1 - Change-point estimation using new Minimum Message Length approximations

AU - Fitzgibbon, Leigh J

AU - Dowe, David L

AU - Allison, Lloyd

N1 - Conference code: 7th

PY - 2002

Y1 - 2002

N2 - This paper investigates the coding of change-points in the information-theoretic Minimum Message Length (MML) framework. Change-point coding regions affect model selection and parameter estimation in problems such as time series segmentation and decision trees. The Minimum Message Length (MML) and Minimum Description Length (MDL78) approaches to change-point problems have been shown to perform well by several authors. In this paper we compare some published MML and MDL78 methods and introduce some new MML approximations called ‘MMLDc’ and ‘MMLDF’. These new approximations are empirically compared with Strict MML (SMML), Fairly Strict MML (FSMML), MML68, the Minimum Expected Kullback-Leibler Distance (MEKLD) loss function and MDL78 on a tractable binomial change-point problem.

AB - This paper investigates the coding of change-points in the information-theoretic Minimum Message Length (MML) framework. Change-point coding regions affect model selection and parameter estimation in problems such as time series segmentation and decision trees. The Minimum Message Length (MML) and Minimum Description Length (MDL78) approaches to change-point problems have been shown to perform well by several authors. In this paper we compare some published MML and MDL78 methods and introduce some new MML approximations called ‘MMLDc’ and ‘MMLDF’. These new approximations are empirically compared with Strict MML (SMML), Fairly Strict MML (FSMML), MML68, the Minimum Expected Kullback-Leibler Distance (MEKLD) loss function and MDL78 on a tractable binomial change-point problem.

U2 - 10.1007/3-540-45683-X_28

DO - 10.1007/3-540-45683-X_28

M3 - Conference Paper

SN - 3540440380

T3 - Lecture Notes in Computer Science

SP - 244

EP - 254

BT - PRICAI 2002: Trends in Artificial Intelligence

A2 - Ishizuka, Mitsuru

A2 - Sattar, Abdul

PB - Springer

CY - Berlin Germany

T2 - Pacific Rim International Conference on Artificial Intelligence 2002

Y2 - 18 August 2002 through 22 August 2002

ER -

Fitzgibbon LJ, Dowe DL, Allison L. Change-point estimation using new Minimum Message Length approximations. In Ishizuka M, Sattar A, editors, PRICAI 2002: Trends in Artificial Intelligence: 7th Pacific Rim International Conference on Artificial Intelligence Tokyo, Japan, August 18-22, 2002 Proceedings. Berlin Germany: Springer. 2002. p. 244-254. (Lecture Notes in Computer Science). https://doi.org/10.1007/3-540-45683-X_28