Bayesian information reward

Lucas R Hope; Kevin B Korb

doi:10.1007/3-540-36187-1_24

Bayesian information reward

Lucas R Hope, Kevin B Korb

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

8 Citations (Scopus)

Abstract

We generalize an information-based reward function, introduced by Good (1952), for use with machine learners of classification functions. We discuss the advantages of our function over predictive accuracy and the metric of Kononenko and Bratko (1991). We examine the use of information reward to evaluate popular machine learning algorithms (e.g., C5.0, Naive Bayes, CaMML) using UCI archive datasets, finding that the assessment implied by predictive accuracy is often reversed when using information reward.

Original language	English
Title of host publication	AI 2002: Advances in Artificial Intelligence
Subtitle of host publication	15th Australian Joint Conference on Artificial Intelligence Canberra, Australia, December 2-6, 2002 Proceedings
Editors	Bob McKay, John Slaney
Place of Publication	Berlin Germany
Publisher	Springer
Pages	272-283
Number of pages	12
ISBN (Print)	3540001972
DOIs	https://doi.org/10.1007/3-540-36187-1_24
Publication status	Published - 2002
Event	Australasian Joint Conference on Artificial Intelligence 2002 - Canberra, Australia Duration: 2 Dec 2002 → 6 Dec 2002 Conference number: 15th https://link.springer.com/book/10.1007/3-540-36187-1 (Proceedings)

Publication series

Name	Lecture Notes in Computer Science
Publisher	Springer
Volume	2557
ISSN (Print)	0302-9743

Conference

Conference	Australasian Joint Conference on Artificial Intelligence 2002
Abbreviated title	AI 2002
Country/Territory	Australia
City	Canberra
Period	2/12/02 → 6/12/02
Internet address	https://link.springer.com/book/10.1007/3-540-36187-1 (Proceedings)

Access to Document

10.1007/3-540-36187-1_24

Cite this

@inproceedings{deebad0980464abb83cad43a2cbc2603,

title = "Bayesian information reward",

abstract = "We generalize an information-based reward function, introduced by Good (1952), for use with machine learners of classification functions. We discuss the advantages of our function over predictive accuracy and the metric of Kononenko and Bratko (1991). We examine the use of information reward to evaluate popular machine learning algorithms (e.g., C5.0, Naive Bayes, CaMML) using UCI archive datasets, finding that the assessment implied by predictive accuracy is often reversed when using information reward.",

author = "Hope, \{Lucas R\} and Korb, \{Kevin B\}",

year = "2002",

doi = "10.1007/3-540-36187-1\_24",

language = "English",

isbn = "3540001972",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "272--283",

editor = "Bob McKay and John Slaney",

booktitle = "AI 2002: Advances in Artificial Intelligence",

address = "Switzerland",

note = "Australasian Joint Conference on Artificial Intelligence 2002, AI 2002 ; Conference date: 02-12-2002 Through 06-12-2002",

url = "https://link.springer.com/book/10.1007/3-540-36187-1",

}

Hope, LR & Korb, KB 2002, Bayesian information reward. in B McKay & J Slaney (eds), AI 2002: Advances in Artificial Intelligence: 15th Australian Joint Conference on Artificial Intelligence Canberra, Australia, December 2-6, 2002 Proceedings. Lecture Notes in Computer Science, vol. 2557, Springer, Berlin Germany, pp. 272-283, Australasian Joint Conference on Artificial Intelligence 2002, Canberra, Australian Capital Territory, Australia, 2/12/02. https://doi.org/10.1007/3-540-36187-1_24

Bayesian information reward. / Hope, Lucas R; Korb, Kevin B.
AI 2002: Advances in Artificial Intelligence: 15th Australian Joint Conference on Artificial Intelligence Canberra, Australia, December 2-6, 2002 Proceedings. ed. / Bob McKay; John Slaney. Berlin Germany: Springer, 2002. p. 272-283 (Lecture Notes in Computer Science; Vol. 2557).

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

TY - GEN

T1 - Bayesian information reward

AU - Hope, Lucas R

AU - Korb, Kevin B

N1 - Conference code: 15th

PY - 2002

Y1 - 2002

N2 - We generalize an information-based reward function, introduced by Good (1952), for use with machine learners of classification functions. We discuss the advantages of our function over predictive accuracy and the metric of Kononenko and Bratko (1991). We examine the use of information reward to evaluate popular machine learning algorithms (e.g., C5.0, Naive Bayes, CaMML) using UCI archive datasets, finding that the assessment implied by predictive accuracy is often reversed when using information reward.

AB - We generalize an information-based reward function, introduced by Good (1952), for use with machine learners of classification functions. We discuss the advantages of our function over predictive accuracy and the metric of Kononenko and Bratko (1991). We examine the use of information reward to evaluate popular machine learning algorithms (e.g., C5.0, Naive Bayes, CaMML) using UCI archive datasets, finding that the assessment implied by predictive accuracy is often reversed when using information reward.

UR - https://www.scopus.com/pages/publications/22944471304

U2 - 10.1007/3-540-36187-1_24

DO - 10.1007/3-540-36187-1_24

M3 - Conference Paper

SN - 3540001972

T3 - Lecture Notes in Computer Science

SP - 272

EP - 283

BT - AI 2002: Advances in Artificial Intelligence

A2 - McKay, Bob

A2 - Slaney, John

PB - Springer

CY - Berlin Germany

T2 - Australasian Joint Conference on Artificial Intelligence 2002

Y2 - 2 December 2002 through 6 December 2002

ER -

Bayesian information reward

Abstract

Publication series

Conference

Access to Document

Other files and links

Cite this