Unified algorithm for undirected discovery of exception rules

Einoshin Suzuki, Jan M. Żytkow

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

19 Citations (Scopus)

Abstract

This paper presents an algorithm that seeks every possible exception rule which violates a common sense rule and satisfies several assumptions of simplicity. Exception rules, which represent systematic deviation from common sense rules, are often found interesting. Discovery of pairs that consist of a common sense rule and an exception rule, resulting from undirected search for unexpected exception rules, was successful in various domains. In the past, however, an exception rule represented a change of conclusion caused by adding an extra condition to the premise of a common sense rule. That approach formalized only one type of exceptions, and failed to represent other types. In order to provide a systematic treatment of exceptions, we categorize exception rules into eleven categories, and we propose a unified algorithm for discovering all of them. Preliminary results on fifteen real-world data sets provide an empirical proof of effectiveness of our algorithm in discovering interesting knowledge. The empirical results also match our theoretical analysis of exceptions, showing that the eleven types can be partitioned in three classes according to the frequency with which they occur in data.

Original languageEnglish
Title of host publicationPrinciples of Data Mining and Knowledge Discovery - 4th European Conference, PKDD 2000, Proceedings
EditorsDjamel A. Zighed, Jan Komorowski, Jan Zytkow
PublisherSpringer Verlag
Pages169-180
Number of pages12
ISBN (Print)9783540410669
Publication statusPublished - Jan 1 2000
Externally publishedYes
Event4th European Conference on Principles and Practice of Knowledge Discovery in Databases, PKDD 2000 - Lyon, France
Duration: Sep 13 2000Sep 16 2000

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume1910
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other4th European Conference on Principles and Practice of Knowledge Discovery in Databases, PKDD 2000
CountryFrance
CityLyon
Period9/13/009/16/00

Fingerprint

Exception
Violate
Simplicity
Theoretical Analysis
Deviation

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Suzuki, E., & Żytkow, J. M. (2000). Unified algorithm for undirected discovery of exception rules. In D. A. Zighed, J. Komorowski, & J. Zytkow (Eds.), Principles of Data Mining and Knowledge Discovery - 4th European Conference, PKDD 2000, Proceedings (pp. 169-180). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1910). Springer Verlag.

Unified algorithm for undirected discovery of exception rules. / Suzuki, Einoshin; Żytkow, Jan M.

Principles of Data Mining and Knowledge Discovery - 4th European Conference, PKDD 2000, Proceedings. ed. / Djamel A. Zighed; Jan Komorowski; Jan Zytkow. Springer Verlag, 2000. p. 169-180 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1910).

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

Suzuki, E & Żytkow, JM 2000, Unified algorithm for undirected discovery of exception rules. in DA Zighed, J Komorowski & J Zytkow (eds), Principles of Data Mining and Knowledge Discovery - 4th European Conference, PKDD 2000, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1910, Springer Verlag, pp. 169-180, 4th European Conference on Principles and Practice of Knowledge Discovery in Databases, PKDD 2000, Lyon, France, 9/13/00.
Suzuki E, Żytkow JM. Unified algorithm for undirected discovery of exception rules. In Zighed DA, Komorowski J, Zytkow J, editors, Principles of Data Mining and Knowledge Discovery - 4th European Conference, PKDD 2000, Proceedings. Springer Verlag. 2000. p. 169-180. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Suzuki, Einoshin ; Żytkow, Jan M. / Unified algorithm for undirected discovery of exception rules. Principles of Data Mining and Knowledge Discovery - 4th European Conference, PKDD 2000, Proceedings. editor / Djamel A. Zighed ; Jan Komorowski ; Jan Zytkow. Springer Verlag, 2000. pp. 169-180 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{fa0344e1ef2f4418affffeb8cd4de4b5,
title = "Unified algorithm for undirected discovery of exception rules",
abstract = "This paper presents an algorithm that seeks every possible exception rule which violates a common sense rule and satisfies several assumptions of simplicity. Exception rules, which represent systematic deviation from common sense rules, are often found interesting. Discovery of pairs that consist of a common sense rule and an exception rule, resulting from undirected search for unexpected exception rules, was successful in various domains. In the past, however, an exception rule represented a change of conclusion caused by adding an extra condition to the premise of a common sense rule. That approach formalized only one type of exceptions, and failed to represent other types. In order to provide a systematic treatment of exceptions, we categorize exception rules into eleven categories, and we propose a unified algorithm for discovering all of them. Preliminary results on fifteen real-world data sets provide an empirical proof of effectiveness of our algorithm in discovering interesting knowledge. The empirical results also match our theoretical analysis of exceptions, showing that the eleven types can be partitioned in three classes according to the frequency with which they occur in data.",
author = "Einoshin Suzuki and Żytkow, {Jan M.}",
year = "2000",
month = "1",
day = "1",
language = "English",
isbn = "9783540410669",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "169--180",
editor = "Zighed, {Djamel A.} and Jan Komorowski and Jan Zytkow",
booktitle = "Principles of Data Mining and Knowledge Discovery - 4th European Conference, PKDD 2000, Proceedings",
address = "Germany",

}

TY - GEN

T1 - Unified algorithm for undirected discovery of exception rules

AU - Suzuki, Einoshin

AU - Żytkow, Jan M.

PY - 2000/1/1

Y1 - 2000/1/1

N2 - This paper presents an algorithm that seeks every possible exception rule which violates a common sense rule and satisfies several assumptions of simplicity. Exception rules, which represent systematic deviation from common sense rules, are often found interesting. Discovery of pairs that consist of a common sense rule and an exception rule, resulting from undirected search for unexpected exception rules, was successful in various domains. In the past, however, an exception rule represented a change of conclusion caused by adding an extra condition to the premise of a common sense rule. That approach formalized only one type of exceptions, and failed to represent other types. In order to provide a systematic treatment of exceptions, we categorize exception rules into eleven categories, and we propose a unified algorithm for discovering all of them. Preliminary results on fifteen real-world data sets provide an empirical proof of effectiveness of our algorithm in discovering interesting knowledge. The empirical results also match our theoretical analysis of exceptions, showing that the eleven types can be partitioned in three classes according to the frequency with which they occur in data.

AB - This paper presents an algorithm that seeks every possible exception rule which violates a common sense rule and satisfies several assumptions of simplicity. Exception rules, which represent systematic deviation from common sense rules, are often found interesting. Discovery of pairs that consist of a common sense rule and an exception rule, resulting from undirected search for unexpected exception rules, was successful in various domains. In the past, however, an exception rule represented a change of conclusion caused by adding an extra condition to the premise of a common sense rule. That approach formalized only one type of exceptions, and failed to represent other types. In order to provide a systematic treatment of exceptions, we categorize exception rules into eleven categories, and we propose a unified algorithm for discovering all of them. Preliminary results on fifteen real-world data sets provide an empirical proof of effectiveness of our algorithm in discovering interesting knowledge. The empirical results also match our theoretical analysis of exceptions, showing that the eleven types can be partitioned in three classes according to the frequency with which they occur in data.

UR - http://www.scopus.com/inward/record.url?scp=84974683992&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84974683992&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84974683992

SN - 9783540410669

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 169

EP - 180

BT - Principles of Data Mining and Knowledge Discovery - 4th European Conference, PKDD 2000, Proceedings

A2 - Zighed, Djamel A.

A2 - Komorowski, Jan

A2 - Zytkow, Jan

PB - Springer Verlag

ER -