符合SCORM標準之學習資源庫的管理機制之研究

Abstract

隨著網際網路的發展，網路學習(e-Learning)也越來越普及。為了促進學習資源在不同網路學習系統間的分享與再利用，近年來有許多國際性組織提出了各種格式的標準，其中最被廣泛應用的是SCORM。另外，在e-learning系統中的學習資源通常都存放在資源庫(Learning Object Repository (LOR) )中，而當資源庫中存放著大量物件時，隨即會面臨到大量物件的管理問題。因此在本篇論文中，我們提出了一個階層式的管理機制(Level-wise Content Management System (LCMS) )來有效地管理符合SCORM標準的學習資源庫。LCMS的流程可分為“建構”與“搜尋”兩大部份。在建構階段(Constructing Phase)，我們先運用SCORM標準中所提供的資訊，將學習資源轉換成一個樹狀架構。接著考慮到SCORM中的詮釋性資料(Metadata)對一般人的複雜度，另外提出了一個方式來輔助使用者來加強學習資源中各學習物件的詮釋性資訊。而後藉由分群的技術，我們針對資源庫中的學習物件建立了一個多層有向非環圖，稱為Level-wise Content Clustering Graph (LCCG)來儲存物件的資訊以及學習物件間的關聯。在搜尋階段(Searching Phase)，提出了一個搜尋機制以利用已建立的LCCG找出使用者想要的學習物件。除此之外，考量到使用者在下搜尋關鍵字時的難處，在此亦基於LCCG提出了一個方式來輔助使用者改善搜尋用詞以在學習資源庫中找出相關的物件。最後，我們實作了一個雛形系統並進行了一些實驗。由實驗結果可知，LCMS的確能有效地管理符合SCORM標準的學習資源庫。

With rapid development of the Internet, e-learning system has become more and more popular. Currently, to solve the issue of sharing and reusing of learning contents in different e-learning systems, several standards formats have been proposed by international organizations in recent years, and Sharable Content Object Reference Model (SCORM) is the most popular one among existing international standards. In e-learning system, learning contents are usually stored in database, called Learning Object Repository (LOR). In LOR, a huge amount of SCORM learning contents including associated learning objects will result in the issues of management over wired/wireless environment. Therefore, in this thesis, we propose a management approach, called Level-wise Content Management Scheme (LCMS), to efficiently maintain, search, and retrieve the learning contents in SCORM compliant LOR. The LCMS includes two phases: Constructing Phase and Searching Phase. In Constructing Phase, we first transform the content tree (CT) from the SCORM content package to represent each learning materials. Then, considering about the difficulty of giving learning objects useful metadata, an information enhancing module is proposed to assist users in enhancing the meta-information of content trees. Afterward, a multistage graph as Directed Acyclic Graph (DAG) with relationships among learning objects, called Level-wise Content Clustering Graph (LCCG), will be created by applying incremental clustering techniques. In Searching phase, based on the LCCG, we propose a searching strategy to traverse the LCCG for retrieving the desired learning objects. Besides, the short query problem is also one of our concerns. In general, while users want to search desired learning contents, they usually make rough queries. But this kind of queries often results in a lot of irrelevant searching results. So a query expansion method is also proposed to assist users in refining their queries and searching more specific learning objects from a LOR. Finally, for evaluating the performance, a web-based system has been implemented and some experiments also have been done. The experimental results show that our LCMS is efficient and workable to manage the SCORM compliant learning objects.