Abstract:

In Case-Based Reasoning systems, the election of the appropriate cases for the case base and the design of an accurate similarity measure are crucial issues. Case-Base Maintenance techniques provide a way to improve the case base mainly by identifying and deleting the cases that produce noise in a system. In this paper we present a novel method for identifying possible future problems with low confidence solutions based on evolutionary techniques. We call these problems dubious future problems. Moreover, some experiments are provided for illustrating the usage of the method.

• Tagged
• XML
• BibTex