Knowledge Engineering Process for a Rapid Prototyping of Inductive Expert System

The main characteristic of current expert systems is the separation of a knowledge base that may be changed from one application to another from the inference engine that still remains the same across applications. The delay in the development of many expert systems is due to the difficulty in acquiring and eliciting knowledge from the human domain experts. The concept of inductive expert system is thus been devised to overcome such bottleneck by incorporating automatic knowledge acquisition module in the system. According to this new concept, knowledge can now be induced or learned in an automatic way from archived databases that are normally available in most organizations. In this paper, we propose an architecture of the inductive expert system that includes the knowledge engine part to automatically forming expert rules from the stored data. We explain the automatic knowledge creation technique through a simple running example, then followed by a real application. We also provide our Prolog source code in appendices for knowledge engineers to apply our technique as a rapid prototyping of their own expert systems.