循環合作式本體論建置系統

Abstract

在數位學習領域中，智慧型數位學習系統應用本體論來提供學習者適性化教學的引導及有效地管理數位學習教材。但是新的概念及相關的子概念會伴隨著新教材出現，使得建置教材管理系統的本體論變得非常困難。然而，如何透過社群提供知識來建置社群共識的本體論變成一個有趣且具挑戰性的議題;我們定義此問題為本體論結晶。由於新的概念會伴隨著教材不斷的產生，需要漸進地建置社群共識的本體論來描述新概念的教材，因此提出循環式收斂機制來減低建置本體論的時間與人力。在每個循環中，社群提供知識來合作建置本體論，並透過收斂機制來評斷社群共識的層級。我們提出循環本體論建置系統來解決知識結晶問題，在系統中，我們提出Wiki-like本體論編輯器作為合作貢獻知識的平台，Questionnaire-based crystallizer利用Delphi-like的方法循環地收斂社群知識。由我們實驗得知，使用ICOC系統能建構出更高層級的社群共識的本體論。

In e-Learning domain, ontology of subject knowledge is applied for intelligent e-learning systems to provide learners with adaptive learning guidance and efficient learning content management. However, the ontology construction for the learning content management system is difficult because new concepts or related sub-concepts are always required along with the new contents. Therefore, “how do we construct the ontology via community to achieve social agreement” becomes a challenging and interesting issue and we define it as the Ontology Crystallization problem. Since making a social agreement its corresponding new version ontology is needed to represent new content for learning content management system, the idea of iterative-based convergence process is proposed to reduce the effort of construction. In each iteration, community members can contribute their knowledge collaboratively to incrementally construct the ontology. Finally, the degree of social agreement can be estimated for the ontology. With the ideas above, we propose an Iterative, Collaborative Ontology Construction (ICOC) scheme, where the Wiki-like ontology editor is proposed as collaborative knowledge contribution platform. Next, a Questionnaire-based crystallizer is proposed as the iterative convergence process with Delphi-like method. The experimental result shows that the ontology constructed by ICOC scheme can achieve higher degree of social agreement.