Automatic rule generation by the transformation of Expert's Diagram: LIFT

To enhance the efficiency of revealing and refining an expert's knowledge, the Expert's Diagram approach is proposed. The Expert's Diagram proposed in this paper is specifically designed for rule-based consulting systems. Using the Expert's Diagram approach, the knowledge acquisition system LIFT is developed, which transforms the Expert's Diagram automatically to rules in the syntax of the shell SKI 2 that is developed for tax consulting-purposes. For the transformation, either the conclusion-directed approach or the condition-directed approach can be applied. The role of LIFT can be generalized to some extent to adapt to the changes in the Expert's Diagram and target shells. According to our experience in the acquisition of Korean corporate tax knowledge, experts could reveal their knowledge effectively using the Expert's Diagram after a short period of training. Thus the rules could be generated automatically by LIFT without the aid of knowledge engineers.