Bridging the gap between business objectives and parameters of data mining algorithms

Data mining is being touted as having potential for discovering information that will contribute toward resolving business problems or creating business opportunities. However, before this potential can be realized, one needs to be able to relate the strategic objectives of a business to the use of...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Decision Support Systems Jg. 21; H. 1; S. 3 - 15
Hauptverfasser: Özden Gür Ali, F., Wallace, William A.
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Amsterdam Elsevier B.V 01.09.1997
Elsevier Sequoia S.A
Schlagworte:
Algorithm parameters
Algorithms
Business objectives
Data mining
Decision support systems
Machine learning
Managerial decision making
Probabilistic inductive learning
Studies
Probabilistic inductive learning
Algorithm parameters
Data mining
Managerial decision making
Business objectives
Machine learning
ISSN:0167-9236, 1873-5797
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Data mining is being touted as having potential for discovering information that will contribute toward resolving business problems or creating business opportunities. However, before this potential can be realized, one needs to be able to relate the strategic objectives of a business to the use of the data mining technology. In this paper we focus on one aspect of that requirement, the translation of managerial goals into parameters of the algorithm being used to analyze the data. The first step in the process is to define a mapping from managerial goals to the performance measures of the algorithm. We then propose and apply an experimental approach using a classification algorithm, Probabilistic Inductive Learning, to develop a mapping between the performance measures and the parameters of the algorithm. The experimental results are analyzed and guidelines are given for setting the bias of the algorithm to meet managerial goals. We conclude by recommending that an experimental analysis similar to what we proposed be conducted for all data mining algorithms and provided along with documentation as an aid to the user.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
content type line 14
ObjectType-Article-2
ObjectType-Feature-1
content type line 23
ISSN:0167-9236
1873-5797
DOI:10.1016/S0167-9236(97)00010-9