| 研究生: |
陳品蓉 Chen, Pin-Rong |
|---|---|
| 論文名稱: |
結合診斷分析之即時學習異常預測方法與技術開發 Development of Method and Technology for Real-Time Learning Anomaly Prediction Integrated with Diagnostic Analysis |
| 指導教授: |
陳裕民
Chen, Yuh-Min |
| 共同指導: |
朱慧娟
Chu, Hui-Chuan |
| 學位類別: |
碩士 Master |
| 系所名稱: |
電機資訊學院 - 製造資訊與系統研究所 Institute of Manufacturing Information and Systems |
| 論文出版年: | 2026 |
| 畢業學年度: | 113 |
| 語文別: | 中文 |
| 論文頁數: | 92 |
| 中文關鍵詞: | 學習分析 、學習異常預測 、診斷分析 、可解釋人工智慧 |
| 外文關鍵詞: | learning analytics, learning anomaly prediction, diagnostic analysis, explainable artificial intelligence |
| 相關次數: | 點閱:10 下載:0 |
| 分享至: |
| 查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
隨著數位學習平台的普及,學習者於學習歷程中所產生之行為、情意與表現資料日益豐富,使得透過資料分析即時掌握學習狀態成為可能。然而,現行多數學習成效分析方法仍偏重於學習歷程結束後之總結性評量,較難即時辨識學習異常並提供具體的輔導依據,亦缺乏對異常成因之系統化解釋。
因此本研究以形成性評量為核心,提出一套結合診斷分析之即時學習異常預測方法與技術架構,目的在於學習歷程進行中,同步進行學習異常預測與潛在成因分析。研究整合多元學習歷程資料,涵蓋學習者之數位行為足跡、情意狀態、疲勞程度等動態數據,並納入先備能力與人格特質等靜態因素,建立多模態學習分析資料集,以供後續模型訓練與異常成因之關聯性探究。在模型設計上,本研究提出一種兼顧預測效能與可解釋性之輕量化模型架構,透過具正權重約束之線性轉換層與注意力機制,使模型於進行預測的同時,可輸出各影響因素對學習異常之作用方向與相對影響程度,作為後續診斷分析之依據。
實驗結果顯示,所提出之模型於學習情意、學習態度與步驟學習成效等預測任務中,均優於線性回歸、支持向量回歸、XGBoost 與 TabNet 等對照模型,展現良好之預測效能與穩定性。而診斷分析亦顯示,不同學習成效與學習狀態群體在影響因素結構上存在差異,突顯個人特質、情緒狀態與學習歷程條件於學習異常形成過程中的交互作用。
Recent analyses of digital learning performance have primarily focused on summative assessment, which makes it challenging to identify learning anomalies and provide timely instructional support. To address this limitation, this study proposes a real-time learning anomaly prediction framework that integrates diagnostic analysis grounded in formative assessment principles. The proposed framework simultaneously predicts learning anomalies and explores their underlying causes by leveraging multimodal data, including digital behaviors, emotional states, fatigue, prior ability, and personality traits.
The proposed model architecture is lightweight and interpretable, incorporating constrained linear transformations alongside attention mechanisms. This design not only enables the model to produce accurate predictions but also offers diagnostic insights into the direction and relative importance of various influencing factors. Experimental results demonstrate that this approach surpasses other inherently interpretable models, such as linear regression, support vector regression, XGBoost, and TabNet, in predicting learning affect, attitude, and performance, achieving greater overall accuracy and stability.
The model's diagnostic analyses reveal unique influence patterns across different learner groups, underscoring how personal traits, emotions, and learning conditions impact performance. This understanding improves our awareness of key factors, making the approach effective for accurate predictions and valuable for informing educational strategies and interventions.
中文參考文獻
均一教育平台,均一教育平台,2024。
辛俊德,全面品質管理在學校管理之實踐,中正教育研究,5卷,1期,91–123,2006。
國立臺中教育大學,113縣市學生學力檢測正式施測題本(國語文三年級),2025。
張存真, 游錦雲,「兒童五大人格特質量表」之信效度及測量不變性:探索性結構方程模式取向,測驗學刊,67卷,3期,2020。
教育部,因材網:適性化學習平台,2024。
黃惠婷,PDCA循環應用於幼兒園教學品質管理系統之研究,淡江大學PhD Thesis,2020。
歐滄和,敎育測驗與評量,心理出版社股份有限公司,Taibei Shi,2002。
學習吧,學習吧平台操作指南,2024。
醍摩豆,2024 年度專刊,2024。
鍾安靜, 陸昱任,以形成性評量為主體的課室評量新趨勢,教師天地,3–12,2014。
簡紅珠,教育大辭書,國家教育研究院,新北市,2000。
簡茂發,多元化評量之理念與方法,啟發多元智能論文集 (頁 189-197). 臺北: 國立臺灣師範大學. 西文部分,1999。
均一教育平台,均一教育平台,2024。
辛俊德,全面品質管理在學校管理之實踐,中正教育研究,5卷,1期,91–123,2006。
國立臺中教育大學,113縣市學生學力檢測正式施測題本(國語文三年級),2025。
張存真, 游錦雲,「兒童五大人格特質量表」之信效度及測量不變性:探索性結構方程模式取向,測驗學刊,67卷,3期,2020。
教育部,因材網:適性化學習平台,2024。
黃惠婷,PDCA循環應用於幼兒園教學品質管理系統之研究,淡江大學PhD Thesis,2020。
歐滄和,敎育測驗與評量,心理出版社股份有限公司,Taibei Shi,2002。
學習吧,學習吧平台操作指南,2024。
醍摩豆,2024 年度專刊,2024。
鍾安靜, 陸昱任,以形成性評量為主體的課室評量新趨勢,教師天地,3–12,2014。
簡紅珠,教育大辭書,國家教育研究院,新北市,2000。
簡茂發,多元化評量之理念與方法,啟發多元智能論文集 (頁 189-197). 臺北: 國立臺灣師範大學. 西文部分,1999。
英文參考文獻
Arik, S.O., Pfister, T.,TabNet: Attentive Interpretable Tabular Learning,2020。
Baltrusaitis, T., Zadeh, A., Lim, Y.C., Morency, L.-P.,OpenFace 2.0: Facial Behavior Analysis Toolkit,2018 13th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2018),59–66,2018。
Bandura, A.,Social foundations of thought and action: a social cognitive theory,Prentice-Hall,Englewood Cliffs, NJ,1986。
Biggs, J.B., Tang, C.S.,Teaching for quality learning at university: what the student does,McGraw-Hill/Society for Research into Higher Education/Open University Press,Maidenhead,2011。
Boekaerts, M., Corno, L.,Self‐Regulation in the Classroom: A Perspective on Assessment and Intervention,Applied Psychology,54卷,2期,199–231,2005。
Broadbent, J., Poon, W.L.,Self-regulated learning strategies & academic achievement in online higher education learning environments: A systematic review,The Internet and Higher Education,1–13,2015。
Chen, T., Guestrin, C.,XGBoost: A Scalable Tree Boosting System,CoRR,2016。
Cowie, B., Bell, B.,A Model of Formative Assessment in Science Education,Assessment in Education: Principles, Policy & Practice,6卷,1期,101–116,1999。
Darkwa, B.F., Antwi, S.,From Classroom to Online: Comparing the Effectiveness and Student Academic Performance of Classroom Learning and Online Learning,OALib,08卷,07期,1–22,2021。
Deming, W.E.,The New Economics for Industry, Government, Education,MIT Press,Cambridge, MA,1993。
Drucker, H., Burges, C.J., Kaufman, L., Smola, A., Vapnik, V.,Support vector regression machines,Advances in neural information processing systems,1996。
Ferguson, R.,Learning analytics: drivers, developments and challenges,Int. J. Technol. Enhanc. Learn.,4卷,5/6期,304–317,2012。
Ghoddoosian, R., Galib, M., Athitsos, V.,A Realistic Dataset and Baseline Temporal Model for Early Drowsiness Detection,Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops,0–0,2019。
Google Research,MediaPipe BlazeFace Model Card (Short-Range),2020。
Google Research,MediaPipe Face Landmark Model with Blendshapes (Version 2),2023。
Google Research,Face Landmark Detection Guide,2024。
Guidotti, R., Monreale, A., Ruggieri, S., Turini, F., Giannotti, F., Pedreschi, D.,A Survey of Methods for Explaining Black Box Models,ACM Computing Surveys,51卷,5期,1–42,2019。
Gupta, A., D’Cunha, A., Awasthi, K., Balasubramanian, V.,DAiSEE: Towards User Engagement Recognition in the Wild,arXiv preprint arXiv:1609.01885,2016。
Hasnine, M.N., Nguyen, H.T., Tran, T.T.T., Bui, H.T.T., Akçapınar, G., Ueda, H.,A Real-Time Learning Analytics Dashboard for Automatic Detection of Online Learners’ Affective States,Sensors,23卷,9期,4243,2023。
Hattie, J.,Visible Learning,Routledge,2008。
Ikawati, Y., Al Rasyid, M.U.H., Winarno, I.,Student Behavior Analysis to Detect Learning Styles in Moodle Learning Management System,2020 International Electronics Symposium (IES),501–506,2020。
Kamath, A., Biswas, A., Balasubramanian, V.,A Crowdsourced Approach to Student Engagement Recognition in e-Learning Environments,Proceedings of the IEEE Winter Conference on Applications of Computer Vision (WACV),1–9,2016。
Kart, L., Laney, D.,Emerging Role of the Data Scientist and the Art of Data Science,2012。
LeCun, Y.A., Bottou, L., Orr, G.B., Müller, K.-R.,Efficient BackProp,Neural Networks: Tricks of the Trade,Springer Berlin Heidelberg,Berlin, Heidelberg,9–48,2012。
Lundberg, S.M., Lee, S.-I.,A unified approach to interpreting model predictions,CoRR,2017。
Moen, R.D., Norman, C.L.,Circling back,Quality progress,43卷,11期,22,2010。
Molnar, C.,Interpretable Machine Learning: A Guide for Making Black Box Models Explainable,2025。
Nair, V., Hinton, G.E.,Rectified linear units improve restricted boltzmann machines,Proceedings of the 27th International Conference on International Conference on Machine Learning,807–814,2010。
OECD,OECD Future of Education and Skills 2030: OECD Learning Compass 2030,2019。
OpenAI,Embedding Models: text-embedding-3-small,2024。
Papamitsiou, Z., Economides, A.A.,Learning Analytics and Educational Data Mining in Practice: A Systematic Literature Review of Empirical Evidence,Journal of Educational Technology & Society,17卷,4期,49–64,2014。
Pekrun, R.,The Control-Value Theory of Achievement Emotions: Assumptions, Corollaries, and Implications for Educational Research and Practice,Educational Psychology Review,18卷,4期,315–341,2006。
Pekrun, R., Lichtenfeld, S., Marsh, H.W., Murayama, K., Goetz, T.,Achievement Emotions and Academic Performance: Longitudinal Models of Reciprocal Effects,Child Development,88卷,5期,1653–1670,2017。
Poropat, A.E.,A meta-analysis of the five-factor model of personality and academic performance.,Psychological Bulletin,135卷,2期,322–338,2009。
Ribeiro, M.T., Singh, S., Guestrin, C.,「Why Should I Trust You?」: Explaining the Predictions of Any Classifier,CoRR,2016。
Škrlj, B., Džeroski, S., Lavrač, N., Petkovič, M.,Feature Importance Estimation with Self-Attention Networks,,2020。
Solanki, N., Mandal, S.,Engagement Analysis Using DAiSEE Dataset,2022 17th International Conference on Control, Automation, Robotics and Vision (ICARCV),223–228,2022。
Susnjak, T., Ramaswami, G.S., Mathrani, A.,Learning analytics dashboard: a tool for providing actionable insights to learners,International Journal of Educational Technology in Higher Education,19卷,1期,12,2022。
Van Der Linden, D., Frese, M., Meijman, T.F.,Mental fatigue and the control of cognitive processes: effects on perseveration and planning,Acta Psychologica,113卷,1期,45–65,2003。
Wilson, D.L.,Asymptotic Properties of Nearest Neighbor Rules Using Edited Data,IEEE Transactions on Systems, Man, and Cybernetics,SMC-2卷,3期,408–421,1972。
Zhang, Y., An, R., Liu, S., Cui, J., Shang, X.,Predicting and Understanding Student Learning Performance Using Multi-Source Sparse Attention Convolutional Neural Networks,IEEE Transactions on Big Data,9卷,1期,118–132,2023。
Zimmerman, B.J.,Attaining Self-Regulation,Handbook of Self-Regulation,Elsevier,13–39,2000。