發展即時動態學習負荷計量機制以建構適性化數位學習狀態診斷模式

Abstract

本研究旨在以動態評量與認知負荷理論為基礎，發展一個數位學習之即時動態學習負荷計量機制，並運用資料探勘技術將學習者依學習風格、自我效能、學習效率進行特性化分組，探討各群組之特徵學習負荷曲線模態與其學習負荷量區間變化率趨勢特性與相對應臨界值，藉此建構適性數位學習狀態診斷模式。研究對象為某科技大學一年級165位學生。 本研究將學習者的個人學習特色納入適性學習機制的設計，並建立出真正即時動態、不受干擾的學習狀態診斷模式，透過使用者導向的雙向互動介面，使得學習者與教學者雙方都能夠於學習歷程中，持續性即時了解學習狀態，進而對系統提出最適化學習路徑做出最佳的學習決策。此成果可運用於建構新一代即時動態適性數位學習系統的規劃與開發，改善現行適性數位學習系統的設計缺失，達成以下四點特性︰﹙一﹚兼具即時動態、不受干擾的互動學習；﹙二﹚即時動態、不受干擾的量化學習狀態回饋；﹙三﹚使用者控制的學習決策；﹙四﹚即時動態之使用者導向的互動介面。

The purpose of this research was to construct a diagnostic model of adaptive e-learning condition based on the theories of dynamic assessment and cognitive load. By using data mining technology, the learners were grouped according to the characteristics of their learning style, self-efficacy, and learning efficiency. To construct a diagnostic model of adaptive e-learning condition, the characterized learning load curve patterns were developed by the real-time dynamic quantitative measurement mechanism of learning load, the variation rate tendency and respective critical figures among each group was analyzed. One hundred and sixty-five freshmen students from a university of technology at southern Taiwan participated in this research. The personal learning characteristic was taken into consideration in the adaptive e-learning mechanism design and the real-time dynamic, non-interference diagnostic model of adaptive e-learning condition was established accordingly. The user-friendly interactive interface enabled the learner and instructor to recognize the learning condition continuously in the learning process and assist a learner in making the optimal learning decision according to the adaptive learning path suggested by the e-learning system. The results of the study can be applied in planning and developing the new generation real-time dynamic adaptive e-learning system in order to improve the defects of existing one and achieve the following purposes: (1) real-time dynamic, non-interference interactive learning, (2) real-time dynamic, non-interference quantitative learning condition feedback, (3) user-controlled learning decision, (4) real-time dynamic user-friendly interactive interface.