人工智慧在教育上的應用：發展教學影片推薦與決策系統以強化翻轉教學

This study used Python as its development tool. First, this study has chosen webpage design as courses of flipped learning. From the basic level to the advanced one, it records several tutorial videos, with recommendation and decision system for students’ self-learning before flipped classroom. Then, students do group discussion and put it into practice in class to make a complete flipped learning process. As the applicant executed current flipped learning program, a problem was founded after the flipped learning became prevalent. Tutoring videos greatly increased, but the level of contents may have discrepancies with students’ abilities, influencing students’ learning performance. Therefore, flipped learning gradually evolved into a burden for students to watch clips at home. In addition, if students found they had known what they watched, they might have no interest watching latter videos. As a result, to solve problems of the previous project, this study employed recommended clips based on students’ capability. It was not necessary for every student to watch all tutorial videos, with which it could prevent students from browsing clips they have learned to reduce their tiredness. Moreover, this study conducted the experiments to verify the system does help learning efficiency. After carrying out this project, this study achieved the following items. (1) recording the instructional and micro-lecture videos of webpage design; (2)Inviting two experts at least to check the key knowledge for each video nodes, and to establish the evaluation items corresponding to the videos; (3)Using python to develop the recommendation and decision system of the instructional videos; (4)Employing examples to train the decision tree; (5) Conducting an instructional experiment to make sure the students in the experimental group have good learning effectiveness when they use the instructional video recommendation system in the flipped teaching; (6)Conducting another experiment to provide the students in the control group with all videos without recommendation in the conventional flipped teaching; It was found that the learning effectiveness of the students watching fewer instructional videos in the experimental group achieved as well as that of those watching complete instructional videos in the control group. (7) The self-efficacy of the students in the experimental had been significantly improved; (8) The cognitive loads of the students in the experimental was lower than that of those in the control group.