The Study on Developing a Learning Emotion Recognition and Emotion Transferring Model during Video-based Programming Learning

• Main Authors: Yu-Fang Lin, 林裕芳
• Other Authors: 許志義
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/9vj8p9

碩士 === 國立中興大學 === 資訊管理學系所 === 106 === Traditional classroom teaching and online learning require evaluation to confirm the effectiveness of student learning. Usually used evaluate methods are summative evaluation and formative evaluation. By observing the facial expressions of students, teacher feedback can be provided during the teaching process to implement formative evaluation. The study mentioned that the use of cognitive affective emotions is more significant than basic emotions in learning. Six kinds of cognitive affective emotions are frustration, confusion, boredom, delight, flow and surprise. Therefore, this thesis renames the above six emotions and neutral emotions as “learning emotions”. This study uses facial expressions to detect students'' emotions in learning process and convert them into emotional transfer paths. The emotional transfer path is used to evaluate the students'' learning outcomes and feedback to the teacher. It is judged whether the students understand the content of the class and adjust the teaching strategies. First, we build the learning emotion recognition model. The learning emotions database is used as the training data of the model. The classifier is support vector machine. The input of the classifier is the feature values by the facial feature points, and the feature selection is use genetic algorithm. The average accuracy of the model is 85.84%. Secondly, collect and mark the students'' facial expression lable in the video based learning, and build the transfer path model for students'' emotions. Inductive of high and low prior knowledge and good or poor learning outcomes. The research results show that: (1) In Video-based learning, confusion to flow, boredom to flow, flow to boredom and flow to confusion were significant. (2) Learner with high prior knowledge and good learning outcome, flow to boredom, boredom to flow, and flow to confusion transitions were significant. (3) Learner with high prior knowledge and poor learning outcome, flow to boredom, boredom to flow were significant. (4) Learner with low prior knowledge and good learning outcome, flow to boredom and boredom to flow were significant. (5) Learner with low prior knowledge and poor learning outcome, flow to boredom and boredom to flow were significant. Finally, based on the above results, the learning effectiveness evaluation model is established through different prior knowledge and different learning outcomes, and different instructional strategies are obtained. The instructional strategies can be feedback to the teacher to adjust the teaching strategy in real time.