Adapting classification rule induction to subgroup discovery

N Lavrac; PA Flach; B Kavsek; L Todorovski

Adapting classification rule induction to subgroup discovery

N Lavrac, PA Flach, B Kavsek, L Todorovski

Research output: Chapter in Book/Report/Conference proceeding › Conference Contribution (Conference Proceeding)

21 Citations (Scopus)

Abstract

Rule learning is typically used for solving classification and prediction tasks. However, learning of classification rules can be adapted also to subgroup discovery. This paper shows how this can be achieved by modifying the covering algorithm and the search heuristic, performing probabilistic classification of instances, and using an appropriate measure for evaluating the results of subgroup discovery. Experimental evaluation of the CN2-SD subgroup discovery algorithm on 17 UCI data sets demonstrates substantial reduction of the number of induced rules, increased rule coverage and rule significance, as well as slight improvements in terms of the area under the ROC curve.

Translated title of the contribution	Adapting classification rule induction to subgroup discovery
Original language	English
Title of host publication	Unknown
Editors	V. Kumar et al.
Publisher	IEEE Computer Society
Pages	266 - 273
Number of pages	7
ISBN (Print)	0769517544
Publication status	Published - Dec 2002

Bibliographical note

Conference Proceedings/Title of Journal: Proceedings of the 2002 IEEE International Conference on Data Mining

Access to Document

http://www.cs.bris.ac.uk/Publications/pub_info.jsp?id=1000689

Handle.net

Persistent link

Cite this

@inproceedings{0458a233d4be4cd599836afb20b6b006,

title = "Adapting classification rule induction to subgroup discovery",

abstract = "Rule learning is typically used for solving classification and prediction tasks. However, learning of classification rules can be adapted also to subgroup discovery. This paper shows how this can be achieved by modifying the covering algorithm and the search heuristic, performing probabilistic classification of instances, and using an appropriate measure for evaluating the results of subgroup discovery. Experimental evaluation of the CN2-SD subgroup discovery algorithm on 17 UCI data sets demonstrates substantial reduction of the number of induced rules, increased rule coverage and rule significance, as well as slight improvements in terms of the area under the ROC curve.",

author = "N Lavrac and PA Flach and B Kavsek and L Todorovski",

note = "Conference Proceedings/Title of Journal: Proceedings of the 2002 IEEE International Conference on Data Mining",

year = "2002",

month = dec,

language = "English",

isbn = "0769517544",

pages = "266 -- 273",

editor = "al., \{V. Kumar et\}",

booktitle = "Unknown",

publisher = "IEEE Computer Society",

address = "United States",

}

TY - GEN

T1 - Adapting classification rule induction to subgroup discovery

AU - Lavrac, N

AU - Flach, PA

AU - Kavsek, B

AU - Todorovski, L

N1 - Conference Proceedings/Title of Journal: Proceedings of the 2002 IEEE International Conference on Data Mining

PY - 2002/12

Y1 - 2002/12

N2 - Rule learning is typically used for solving classification and prediction tasks. However, learning of classification rules can be adapted also to subgroup discovery. This paper shows how this can be achieved by modifying the covering algorithm and the search heuristic, performing probabilistic classification of instances, and using an appropriate measure for evaluating the results of subgroup discovery. Experimental evaluation of the CN2-SD subgroup discovery algorithm on 17 UCI data sets demonstrates substantial reduction of the number of induced rules, increased rule coverage and rule significance, as well as slight improvements in terms of the area under the ROC curve.

AB - Rule learning is typically used for solving classification and prediction tasks. However, learning of classification rules can be adapted also to subgroup discovery. This paper shows how this can be achieved by modifying the covering algorithm and the search heuristic, performing probabilistic classification of instances, and using an appropriate measure for evaluating the results of subgroup discovery. Experimental evaluation of the CN2-SD subgroup discovery algorithm on 17 UCI data sets demonstrates substantial reduction of the number of induced rules, increased rule coverage and rule significance, as well as slight improvements in terms of the area under the ROC curve.

M3 - Conference Contribution (Conference Proceeding)

SN - 0769517544

SP - 266

EP - 273

BT - Unknown

A2 - al., V. Kumar et

PB - IEEE Computer Society

ER -

Adapting classification rule induction to subgroup discovery

Abstract

Bibliographical note

Access to Document

Handle.net

Fingerprint

Cite this