Aul, Vani, Sauer, Christian and Roth-Berghofer, Thomas (2013) Knowledge maintenance in myCBR. In: 18th UKCBR 2013 Workshop, 10 Dec 2012, Cambridge, UK.

Preview

PDF Knowledge Maintenance in myCBR_CRC.pdf - Accepted Version Download (467kB) | Preview

Abstract

CBR systems, being knowledge based systems, process knowledge. Due to changes in the environment a CBR system's knowledge model can become outdated, thus creating a need for constant maintenance of said knowledge model. In this paper, we describe an implementation of (semi-)automatic knowledge maintenance of two of the four knowledge containers of CBR systems, especially case base maintenance and maintenance of similarity measures within the CBR system development SDK myCBR. We describe our approach to create, elicit and manage quality measures that are used to trigger maintenance actions if the quality measures fall below defined thresholds, indicating a declining efficiency/accuracy of a case base or particular similarity measure. We further detail on the implementation of our approach into myCBR Workbench to enable a knowledge engineer to incorporate the notion of maintenance already at the design stage of a CBR system. The approach relies on the notion of maintenance attributes to be able to measure the quality of case bases and similarity measures. Initial experiments using the newly introduced quality measurement attributes indicate that our approach is promising.

Item Type:	Conference or Workshop Item (Paper)
Subjects:	Computing
Depositing User:	Christian Sauer
Date Deposited:	26 May 2016 13:22
Last Modified:	04 Nov 2024 11:02
URI:	https://repository.uwl.ac.uk/id/eprint/2181

Downloads

Actions (login required)

View Item