Integrated knowledge acquisition from text, previously solved cases, and expert memories

Abstract

Within the model-based knowledge engineering framework, an integrated knowledge acquisition method was developed for a complex real-world domain with different traces of expertise. By having an expert constructively explain the previously solved cases with more general information from other traces of expertise (text, expert memories) a model-centered knowledge base is constructed. The proposed method allows for an early knowledge verification where the relevance, sufficiency, redundancy, and consistency of knowledge are already assessed at an informal level. The early knowledge verification efficiently prepares the consecutive knowledge formalization. Through a cognitively adequate model of expertise and the explanation-oriented knowledge elicitation procedures, user friendly second generation expert systems may be developed.