Knowledge realization and transformation in KRISP

Abstract

KRISP is a system written in Prolog for the representation and inference of complex and ill-structured knowledge. Its logic programming approach makes knowledge transformation from experts' knowledge to a executable knowledge base easier. It also helps a process called knowledge realization which is the integration of fragmented knowledge into meaningful rules at each transformation stage. KRISP was applied to electric power system diagnosis and it was found that it performed a little better than domain experts and it was also effective in the problem domains where physical data structure can be mapped onto logical data structure directly.