The full text of this article

 

A revised approach to solving the symbolic value partition problem from a viewpoint of roughness of partitions
by Yasuo Kudo
International Journal of Reasoning-based Intelligent Systems (IJRIS), Vol. 4, No. 3, 2012

 

Abstract: In this paper, we discuss a revision of a heuristic algorithm used to optimise symbolic value partitions that was proposed by Min et al. (2008) (RBSVP algorithm) from a viewpoint of the roughness of partitions. Min et al.'s approach for optimising symbolic value partitions was to minimise the number of attribute values used in a given decision table, and the RBSVP algorithm outputs a suboptimal symbolic value partition. However, in some cases, this approach might not contribute to the extraction of useful decision rules from the results of the optimised decision table. In this paper, instead of using the number of attribute values, we introduce the average coverage score that was proposed by Kudo and Murai (2010a) as a criterion of relative reducts based on the roughness of partitions, and we present an example of how the average coverage score is used to select relative reducts in the revised RBSVP algorithm. The results of our experiments indicate that the revised algorithm works well for extracting useful decision rules.

Online publication date: Fri, 16-Nov-2012

 

is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

 
Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

 
Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Reasoning-based Intelligent Systems (IJRIS):
Login with your Inderscience username and password:

 

    Username:        Password:         

Forgotten your password?


 
Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

 
If you still need assistance, please email subs@inderscience.com