On the consistency of information filters for lazy learning algorithms

H Brighton; C Mellish

On the consistency of information filters for lazy learning algorithms

H Brighton^*, C Mellish

^*Corresponding author for this work

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

40 Citations (Scopus)

Abstract

A common practice when filtering a case-base is to employ a filtering scheme that decides which cases to delete, as well as how many cases to delete, such that the storage requirements are minimized and the classification competence is preserved or improved. We introduce an algorithm that rivals the most successful existing algorithm in the average case when filtering 30 classification problems. Neither algorithm consistently outperforms the other, with each performing well on different problems. Consistency over many domains, we argue, is very hard to achieve when deploying a filtering algorithm.

Original language	English
Title of host publication	PRINCIPLES OF DATA MINING AND KNOWLEDGE DISCOVERY
Editors	JM Zytkow, J Rauch
Publisher	SPRINGER-VERLAG BERLIN
Pages	283-288
Number of pages	6
ISBN (Print)	3540664904
Publication status	Published - 1999
Externally published	Yes
Event	3rd European Conference on Principles of Data Mining and Knowledge Discovery in Databases (PKDD 99) - PRAGUE, Czech Republic Duration: 15 Sep 1999 → 18 Sep 1999

Publication series

Name	LECTURE NOTES IN ARTIFICIAL INTELLIGENCE
Publisher	SPRINGER-VERLAG BERLIN
Volume	1704
ISSN (Print)	0302-9743

Conference

Conference	3rd European Conference on Principles of Data Mining and Knowledge Discovery in Databases (PKDD 99)
Country/Territory	Czech Republic
City	PRAGUE
Period	15/09/99 → 18/09/99

Cite this

@inproceedings{aecdc5ddc0d444e9b0ba429fc032a61d,

title = "On the consistency of information filters for lazy learning algorithms",

abstract = "A common practice when filtering a case-base is to employ a filtering scheme that decides which cases to delete, as well as how many cases to delete, such that the storage requirements are minimized and the classification competence is preserved or improved. We introduce an algorithm that rivals the most successful existing algorithm in the average case when filtering 30 classification problems. Neither algorithm consistently outperforms the other, with each performing well on different problems. Consistency over many domains, we argue, is very hard to achieve when deploying a filtering algorithm.",

author = "H Brighton and C Mellish",

year = "1999",

language = "English",

isbn = "3540664904",

series = "LECTURE NOTES IN ARTIFICIAL INTELLIGENCE",

publisher = "SPRINGER-VERLAG BERLIN",

pages = "283--288",

editor = "JM Zytkow and J Rauch",

booktitle = "PRINCIPLES OF DATA MINING AND KNOWLEDGE DISCOVERY",

note = "3rd European Conference on Principles of Data Mining and Knowledge Discovery in Databases (PKDD 99) ; Conference date: 15-09-1999 Through 18-09-1999",

}

Brighton, H & Mellish, C 1999, On the consistency of information filters for lazy learning algorithms. in JM Zytkow & J Rauch (eds), PRINCIPLES OF DATA MINING AND KNOWLEDGE DISCOVERY. LECTURE NOTES IN ARTIFICIAL INTELLIGENCE, vol. 1704, SPRINGER-VERLAG BERLIN, pp. 283-288, 3rd European Conference on Principles of Data Mining and Knowledge Discovery in Databases (PKDD 99), PRAGUE, Czech Republic, 15/09/99.

On the consistency of information filters for lazy learning algorithms. / Brighton, H; Mellish, C.

PRINCIPLES OF DATA MINING AND KNOWLEDGE DISCOVERY. ed. / JM Zytkow; J Rauch. SPRINGER-VERLAG BERLIN, 1999. p. 283-288 (LECTURE NOTES IN ARTIFICIAL INTELLIGENCE; Vol. 1704).

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

TY - GEN

T1 - On the consistency of information filters for lazy learning algorithms

AU - Brighton, H

AU - Mellish, C

PY - 1999

Y1 - 1999

N2 - A common practice when filtering a case-base is to employ a filtering scheme that decides which cases to delete, as well as how many cases to delete, such that the storage requirements are minimized and the classification competence is preserved or improved. We introduce an algorithm that rivals the most successful existing algorithm in the average case when filtering 30 classification problems. Neither algorithm consistently outperforms the other, with each performing well on different problems. Consistency over many domains, we argue, is very hard to achieve when deploying a filtering algorithm.

AB - A common practice when filtering a case-base is to employ a filtering scheme that decides which cases to delete, as well as how many cases to delete, such that the storage requirements are minimized and the classification competence is preserved or improved. We introduce an algorithm that rivals the most successful existing algorithm in the average case when filtering 30 classification problems. Neither algorithm consistently outperforms the other, with each performing well on different problems. Consistency over many domains, we argue, is very hard to achieve when deploying a filtering algorithm.

M3 - Conference contribution

SN - 3540664904

T3 - LECTURE NOTES IN ARTIFICIAL INTELLIGENCE

SP - 283

EP - 288

BT - PRINCIPLES OF DATA MINING AND KNOWLEDGE DISCOVERY

A2 - Zytkow, JM

A2 - Rauch, J

PB - SPRINGER-VERLAG BERLIN

T2 - 3rd European Conference on Principles of Data Mining and Knowledge Discovery in Databases (PKDD 99)

Y2 - 15 September 1999 through 18 September 1999

ER -

On the consistency of information filters for lazy learning algorithms

Abstract

Publication series

Conference

Fingerprint

Cite this