Epistemological problems in the development of intelligent information systems for industrial applications

Expert knowledge should not be thought of as consisting of static structures that can be neatly divided into rules and facts, then captured by some knowledge representation formalism. Otherwise, knowledge acquisition (KA) would be thought of as a technical process that translates human meanings into symbols that can then be used by a computer to reconstruct specific forms of intelligent human behavior. KA is intended to provide an unambiguous description of the contents of the knowledge-based component of a knowledge system (KS). The quality of KA is critical in building successful KS applications. Researchers have inquired into some of the principal problems of KA as well as explored some of the epistemological issues relevant to the elicitation of knowledge from experts. As a result of this research, a strategy has been developed that is found to be effective for dealing with the problems of capturing embedded implicit knowledge. This KA strategy involves a 4-step procedure - orientation, dialogue, case analysis, and normalization - conducted with 2 KEs and the expert.