現有建構在全球資訊網環境下的遠距教學系統，學生之間因為缺乏理想的即時面對面互動環境，容易導致學生在個別學習時缺乏同儕支援及同儕壓力；而面對網站上龐大的學習歷程，教師亦無法有效地加以處理、適時地觀察學習狀態並擬定教學策略來提升學習效能。 為了增加網路學習環境下的互動性、發揮群體學習環境下的同儕互助以及同儕壓力，並助教師分析學習歷程並建立學習模型，本論文提出了一個在全球資訊網環境下的群體學習系統，提供學生群體學習的環境與機制，並提供教師對於學習歷程知識探索及學習模型建構之工具，以滿足學生及教師在遠距教學生的需求。 本論文在輔助學生學習上，引用社會學中的群體學習理論、人際溝通網路及角色扮演模型，提供學生在網路群體學習時利用，並且針對異質分組、資源共享、溝通討論及專案合作上設計了各種模組以滿足學生在學習活動上之需求；在輔助教師教學上，則利用資料庫系統記錄學生的學習行為及互動關係，並採用資料尋礦及溝通網路分析技術來幫助教師觀察及分析學生的學習特徵及互動關係，並幫助教師依此尋找影響群體學習成效的因素，預測學習表現 ,提供教學策略決策時所需的資訊，進而有效提升學生在網路環境中之學習效果。 When constructing a web group learning system to foster peer collaboration, teachers must monitor the group’s learning status and encourage groups to learn without face-to-face communication. Therefore, an adaptive model for web group learning must be developed to facilitate peer collaboration and to assist teachers in monitoring the group learning status. This thesis presents a method in which computer science approaches and social science analysis are incorporated to support students learning on Internet, and to assist teachers in identifying the group learning status. By becoming aware of group learning status extracted from group learning behavior and communication, teachers can more understand the relationships between group learning behavior and group learning performance, and the relationships between group communication and group learning performance. Fully integrating data mining techniques, information retrieval techniques, machine learning techniques and social network analysis enables teachers to cope with a large quantity of learning web logs and monitor group learning on the Internet. The experiment results show the significant relationships between group communication and group learning performance and the relationships between member-roles and group learning performance. The results enable teachers to monitor and encourage the group learning on Internet.
