Towards Validation of Case-Based Systems

The present paper is an attempt to apply the authors’ experiences of their work in validation of rule based systems (of. (Knauf, Abel, Jantke ~nd Gonzalez 1998). (Knauf, .lantke, Abd anti Philippow 1997), (Abel, Knauf, and Gonzalez 1996), (Abel Gonzalez 1997b), (Herrmann, .lantkc, and Knauf 1997), and (,lantke, Knauf and Abel 1997),t:.g.) c~e based systems. The objective of this work is both to coil: tip with a framework for validation of c~Lse b&,~ed systems and to attswer I, he question how the knowledge representation anti the problem solving paradigm intltiences the validation technology of an AI systent. It turns out, that the gcnerM steps of test case validation (test cruse generation, test case experimentation, evaluation, validity assessment, and system refinement), which are performed in cycles, can and should be the same for both parazligms, but lhe realization of Ihe partictilar steps scents to be different. Most of the steps can be performed with less expendittire and less hnnmtt support for case based systems than for rule based systems. Gcm.r’,dly, the more zL knowledge representatiott level is explicit and abstract, the more the validation of the .~ystem seems to be complicated.