2964938630

Expert systems are problem solvers for spedalized domains of competence in which effective problem solving normally reąuires human expertise. The transition of expert systems technolog}? from research laboratories to software development eenters highlighted the fact that the ąuality assurance for expert system is a very important issue for most real-word problems. Although the basie verification concepts are shared by software engineering and knowledge engineering, verifieation methods of conventional software are not directly applieable to expert systems and the new, specific methods of verification are required. The main aim of this work is to present our own rule base verification method. In our opinion the decision units coneeption allows us to consider differentverifieationandvalidationissuestogether. Thanks to properties of the decision units we can perform different verification and validation actions during knowledge base development and realization.

Expert syste ms are problem solvers for spedalized domains of competence in which effective problem solving normally reąuires human expertise. The transition of expert systems technolog}? from research laboratories to software development centers highlighted the fact that the ąuality assurance for expert system is a very important issue for most real-word problems. Although the basie verification concepts are shared by software engineering and knowledge engineering, verification methods of conventional software are not directly applicable to expert systems and the new, specific methods of verification are reąuired. The main aimofthisworkis to present our own rule base verification method. In our opinion the decision units conception allows us to consider differentverification and validation issuestogpther. Thanks to properties of the dedsion units we can perform different verification and validation actions during knowledge base development and realization.