在這個資訊科技發達的年代,利用電腦和資料庫來儲存生活上或各領域的知識已經非常普遍。但要將專家腦中的專業知識予以系統化是非常困難的,因此有許多學者們紛紛投入這方面的研究。其中凱利方格法是最常見的知識擷取方法,它是以物件、屬性與連接機制為主要元件,再透過一套擷取流程一步一步將專家的專業知識擷取出來。而EMCUD(Embedded Meaning Capturing and Uncertainty Deciding,簡稱 EMCUD)是一套協助專家提供隱含的專業知識的擷取技術,它提出多資料型態表格和AOT(Attribute Ordering Table)表格來儲存知識與其重要程度順序,讓知識的表達可以更加細膩,並提出OBJECT-CHAIN演算法作為知識的擷取方法,然後經由專家的協助推論出每一條規則中的隱含知識。 本論文提出一套相似度判斷的方法,該方法可從目前獲得的知識中去判斷是否有太過相似的物件組,藉此加以區分,並嘗試將凱利方格法與EMCUD技術作結合,然後將相似度判斷加入至該知識擷取技術中。我們也開發出一套以網頁為平台的知識擷取系統,並邀請了四位國小教師使用本系統進行知識擷取的實驗,實驗的主題為校園植物知識的建構。經由實驗結果顯示相似度判斷演算法能適時的發揮作用,達到區分物件與擴展知識的功效。而本系統的擷取流程亦能反覆的檢查當前的知識並對使用者提出相對應的問題導引。
In recent years, information technology is prospering. The computer and database to store the knowledge of all areas has been very common. It's very difficult to organization of the expertise from expert. Therefore, many scholars are invested at the research of knowledge acquisition. However, the repertory grid is the most common in knowledge acquisition methods. It is based on objects, attributes and linking mechanism as the main component to capture the knowledge from experts step-by-step. The second method, EMCUD(Embedded Meaning Capturing and Uncertainty Deciding), can help experts to capture the implicit knowledge. It proposed an acquisition table and an AOT(Attribute Ordering Table) to store knowledge and order of their importance, and then use OBJECT-CHAIN as a approach of knowledge acquisition. This paper presents a method of judgments similarity, which can judge the similar objects from the knowledge table. We try to combine the repertory grid with EMCUD, and added the method of judgments similarity in the knowledge acquisition system. We have invited four teachers to use the system. The subject of our experiment is to construct the knowledge of plants in campus. The experimental results show that, the algorithm of judgments similarity can distinguish objects and extend the effect.