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.
