| Original language | English |
|---|---|
| Pages (from-to) | 246-266 |
| Journal | Knowledge and Information Systems |
| Volume | 7 |
| Issue number | 2 |
| DOIs | |
| Publication status | Published (in print/issue) - 1 Feb 2005 |
Bibliographical note
Other Details------------------------------------
A critical aspect of any intelligent system is the representation of knowledge. Frequently knowledge is represented by a mapping from a condition space to a decision space. The large databases utilised are often built for purposes other than knowledge discovery and incorporate substantial redundancies, thus resulting in increased complexity in the knowledge discovery process. The importance of this paper is that it proposes a new multi-knowledge approach which, when combined with the Bayes classifier, allows for a higher decision accuracy, particularly in the situation of data sets with a large number of attributes.