應用案例式推理建置學習物件內容註解之研究

Abstract

數位學習是個新興的領域，目前趨勢不論數位教材或數位學習平台都遵循SCORM標準。學員在數位學習平台上學習時，有課程問題無法立即穫得老師的解答，客服人員只能夠協助回答平台操作及開班事宜。而回答課程問題往往須要透過專業領域講師，學員人數及課程數量繁多時愈無法應付學員的問題。 為了要解決數位學習在課程問題回覆方面的困難並且有效管理學習物件的內容，本研究以布魯姆知識認知分類理論定義一個學習物件內容註解語言(Learning Object Content Annotation Language, LOCAL)。本研究並採用案例式推理(Case-Based Reasoning, CBR)來開發課程問題回覆系統。本研究以學習物件內容註解語言為特徵值，將教材內容擷取出來做知識管理，建置課程問題知識庫，並且讓系統可以自動擷取及重組教材內容回覆學員問題。經過本研究初步的分析實作驗證，利用終身學習教材，建立近百個註解語言，可以應用在擷取課程知識上。 本系統提供自然語言即時問題回覆，可做為提供數位學平台課程內容服務的工具。

E-Learning is a burgeoning domain in recent years. The current trend is using SCORM standard for both curriculums and e-Learning platform. However the students could not get lecturers’ immediately responses to those questions about curriculums when they are using the current platform. Because e-Learning call center usually only provides assistances for platform operations and courses’ delivering. Owing to the reason that answering questions about curriculums highly depends on manpower of professional lecturers. In order to resolve the above mentioned problems, and providing effectively management for the contents of Learning Object, this research adopts Bloom’s Taxonomy to define a XML language for courseware content annotation, called Learning Object Content Annotation Language (LOCAL). Besides, this research also introduces Case-Base Reasoning (CBR) to develop the question answering system. It uses metadata in LOCAL as the characteristic value for the construction of courseware contents knowledge base. After preliminary analysis, implementation, and verification in this research, we establish nearly hundred annotation metadata by using of lifetime learning courseware materials, which could be applied to knowledge retrieving of curriculums. The system demonstrate the capability of instant questions answering by Natural Language, it can serve with curriculums content services for an e-Learning platform.