| 摘要: | 隨著網路與資訊科技的持續發展，線上數位學習越來越盛行，學生學習的管道不再侷限於面對面的實體課程。同時，2020受到COVID-19新冠肺炎疫情的影響，使得人們學習需求與學習的行為有了非常大幅度的改變。因疫情的影響使人們需保持社交距離的政策，讓世界各國對線上數位學習的重視程度有明顯的提升。常見的線上數位學習平台有Coursera、Moocs、edX，及本論文研究的BookRoll線上電子書學習平台。
而BookRoll線上電子書學習平台是由日本京都大學所開發的一套電子書閱讀系統，不僅支援行動學習，也提供了書籤(Bookmark)、標記(Marker)、備忘錄(Memo)等相關功能等，其系統能夠記錄學生的所有閱讀行為，藉此蒐集學生完整的學習日誌(Log)。同時能搭配複習系統、概念評量系統、程式評量系統來評估學生的學習狀況。
本論文研究將使用學生於BookRoll線上電子書學習平台上的閱讀行為紀錄、複習、評量系統上的答題紀錄作為數據基礎。透過資料前處理提取特徵出來，並先使用統計方法，來分析學生閱讀行為與學習成效之間的關聯性。接著利用關聯分析來探索閱讀行為與學生學習成效的關係，最後，使用機器學習分類演算法對學生的學習成效進行預測分類，並比較不同的分類演算法的預測成效。;With the continuous development of the Internet and information technology, online digital learning has become more and more popular, and the way students learn is no longer limited to physical courses. Besides, society was affected by the COVID-19 pandemic in 2020, which made people′s learning needs and learning behaviors have changed significantly. Due to the pandemic, social distancing is a new policy that has apparently increased the importance of digital learning around the world. Common online digital learning platforms include Coursera, Moocs, edX, and the BookRoll, the online e-book learning platform studied in this paper.
The BookRoll online e-book learning platform is a reading system developed by Kyoto University in Japan. It not only supports active learning but also provides Bookmark, Marker, Memo and other functions. The system collects students’ complete learning logs by recording all the reading behaviors of students. Moreover, it can be used with a review system, concept evaluation system, and program evaluation system to evaluate students′ learning situations.
Students′ reading behavior records, review records, and answer records on the evaluation system on the BookRoll online e-book learning platform will be applied as the data basis in this research. Features are extracted and statistical methods are used to analyze the correlation between students′ reading behavior and learning outcomes. Next, analyzing the relationship between reading behavior and students′ learning outcomes by the association rule. Lastly, using machine learning classification algorithms to predict and classify students′ learning outcomes and compare the predicted performances of different classification algorithms. |
