著者

大内 東 栗原 正仁 加地 郁夫 Azuma Ohuchi Masahito Kurihara Ikuo Kaji

出版者

一般社団法人 人工知能学会

雑誌

人工知能学会誌 = Journal of Japanese Society for Artificial Intelligence (ISSN:09128085)

巻号頁・発行日

vol.3, no.5, pp.599-606, 1988-09-20

被引用文献数

4

---

The bottleneck in the process of building expert systems is retrieving the appropriate problem solving knowledge from the human expert. Method of developing an understanding of complex situations from systems engineering based on Interpretive Structural Modeling(ISM)is applied to this process. A theoretical framework of the lnteractive Knowledge Structural Modeling(IKSM) is described which interactively developing the knowledge structures by computer asistence. IKSM process is devided into two main phases, embedding phase and analytic phase. The embedding phase is to elicit the knowledge needed to solve analysis problems. Three implication rules are derived and applied to effectively execute the embedding phase. The analytic phase then arranges the knowledge to a hierarchy structure. This process extracts equivalence class, partial ordering, covering, and connected parts from the knowledge. The result is presented by some figure easily understood by the human expert.