Rainer Knauf and Ilka Philippow, Ilmenau Technical University, Germany; Avelino J. Gonzalez, University of Central Florida, USA; Klaus P. Jantke, German Research Center for Artificial Intelligence Ltd., Germany
A methodology for the validation of rule-based expert systems is presented as a 5-step process that has three central themes: (1) creation of a minimal set of test inputs that cover the domain, (2) a Turing Test-like methodology that evaluates the system’s responses to the test inputs and compares it to the responses of human experts, and (3) use the validation results for system improvement. This methodology can be performed in loops. The starting point of each cycle is a rule base and the loop ends up in a (hopefully) better rule base. The first three steps of this process have been published as separate issues in earlier papers by the authors. This paper gives an overview on the entire process and describes the relation between the steps and the system refinement step. In this last step, the rules are modified according to the results of evaluating the test cases. The base of this rule base reconstruction is both a rule-associated validity and the existence of a better rated human solution.