江佳倩
江佳倩
智能式學習支持方法與實現技術研發：以身心障礙學生社會技巧學習為例
Development of Smart Learning Support Method and Enabling Technology: Social Skills Learning of Students with Disabilities as an Example
特殊教育、身心障礙、社會技巧、遊戲式學習、學習支持系統、機器學習、類神經網路
Learning Support System、Game-Based Learning、Disabilities、Machine Learning、Neural Network、Social Skill、Special Education
因材施教、適性化的學習支持是數位學習最具有代表性的議題。由於身心障礙學生具個別差異大的特質，更需要以學習者為中心，提供適性化的教學策略與個別化輔導與支持，以幫助學生解決學習困難、提升學習成效。且身心障礙學生因個人特質，常會有社會互動巧不足、人際狀況不佳等困擾，因此社會技巧學習對於身心障礙學生而言十分重要。
本研究之主要目的在參考資料科學之資料分析(Data Analytic)與最佳化(Optimization)之概念，設計適用於數位遊戲學習之「智能式學習支持方法」，再運用機器學習技術，建構學生模型與適性化輔導決策模型，以及模型之優化機制，以在行為變化多端、複雜的互動與學習過程中，提供即時與適性化的輔導。
本研究於國立成功大學製造資訊與系統研究所「智能商務實驗室」所開發之「遊戲式社會技巧數位學習平台」進行實驗，以檢驗此一「智能式學習支持方法」之可行性與妥適性，以及對身心障礙學生社會技巧提升成效。實驗結果顯示受試者之問題行為出現頻率明顯下降，且正向行為亦有所提升，此外亦有部分學生之電腦操作技巧有所提升。
Teaching students in accordance with their aptitude, and giving their appropriate learning support is the most representative issue of e-learning. In particular, students with disabilities have the characteristics of individual differences. They need appropriate learning support even more. We must provide a learner-centered adaptive assist strategy to help students solve learning difficulties and improve learning outcomes. Because of their personal traits, “deficits in social interaction ability” and “poor interpersonal condition” are common problems among students with disabilities. Therefore, social skills learning is very important for students with disabilities.
The main purpose of this study is to design the Smart Learning Support Method for digital game-based learning based on Data Analysis and Optimization Concept of Data Science. Apply the machine learning technology to establish the student model, assist strategy decision model and the optimization mechanism of the model. Provide instant adaptive support in a complex and interactive process of learning.
The experiment in this study was conducted at the Game-based Social Skills Digital Learning Platform to verify the feasibility and appropriateness of this Smart Learning Support Method. Finally, through a series of experiments, the results showed that the frequency of the subject's inappropriate behavior decreased significantly, and the positive behavior also improved. In addition, some students' computer skills were improved.