情緒感知學習系統融入翻轉教室課程之學習成效評估

Effects of Affective Awareness Learning System in Flipped Classroom Course on Students’ Learning Performance

在翻轉教室中，學生於課堂外的自我學習活動教材多以單向傳遞數位教材為主，然而，因課堂外缺乏與同儕及教師之引導與互動，對較缺乏學習動機的學生而言易形成學習成效低落的情形，此外，不同於課堂中教師可於教學過程中觀察學生之學習情緒，進而提供即時輔助，在課堂外學生所進行之學習活動中，因教師與同儕之互動與輔助更顯缺乏，故學生之學習情緒反饋與即時輔助更顯重要。綜觀上述，本研究以人工智慧深度學習技術為基礎，應用情感運算模型並結合LINE機器人服務開發情緒感知學習系統，讓學生在翻轉教室課堂外進行自我學習時，能夠基於學生之學習狀態及學習情緒提供學習服務，進而提升其於翻轉教室課堂外自我學習活動中之互動性與學習情緒的反饋。為評估本系統融入翻轉教室對於學生學習成效之影響，本研究規劃一準實驗，以傳統翻轉教室導入大學軟體工程課程為控制組，本系統融入翻轉教室導入大學軟體工程課程為實驗組，實驗結果指出，相較於傳統翻轉教室教學方法，本研究所提出之系統融入翻轉教室課程可有效提升學生的學習成效、學習動機及問題解決能力，學生學習動機對於問題解決能力及學習成就有正向影響，且問題解決能力對學習動機與學習成就呈現正向部分中介效果。

In the flipped classroom, students usually conduct self-learning alone out of the class. In this situation, the students usually need appropriate guidance and peer and teacher interactions to support the self-learning. However, the students with low learning motivation may not engage in such online platforms actively and thus affect their learning performance. Therefore, appropriate learning guidance and interactive mechanisms are needed to develop for supporting the students’ self-learning process out of the class in the flipped classroom. Moreover, since the teacher can observe the students’ learning emotion and further provide appropriate learning supports to the students in the class, the teacher cannot provide such supports to the students out of the class in the flipped classroom. Therefore, it is important to the flipped classroom that is how to detect the students’ learning emotion and further provide appropriate learning supports out of the class. Bearing this in mind, this study aims to apply artificial intelligence and affective computing techniques to develop an affective awareness learning system with LINE bot for facilitating students’ self-learning out of class in the flipped classroom. To evaluate the effects of the proposed system in flipped classroom on students’ learning performance, two classes of students were invited to conducted a quasi-experiment in this study. One class served as the control group and the other class served as the experimental group. The control group was supported by traditional flipped learning strategy. The experimental group was supported by flipped learning strategy with the proposed system. The experimental results showed that the students who learned with the proposed approach outperformed those students who learned with traditional flipped learning approach in terms of learning performance, learning motivation, and problem solving ability. Moreover, the students’ learning motivation positively affects problem solving ability and learning achievement. In addition, the students’ problem solving ability showed a positively, partial mediation between learning motivation and learning achievement.