臺灣博碩士論文加值系統

隨著科技進程與網際網路的廣泛運用，線上開放教育資源也同時逐漸興起。大規模開放線上課程（MOOC，Massive Open Online Course）為近期最新的遠端教育指標，特色為開放共用（open access）及可擴張性（scalability）。我國教育部於2014年初發展磨課師（MOOCs）課程推動計畫，透過Coursera、Udacity、Open edX提供的開放源碼，各大專院校也逐漸發展出自己的開放線上課程，目的為打造學生於學習上不受限於時間、空間及地點的學習環境，也希望基於學生廣泛地參與能夠透過這些學習數據能為提供更多學生的學習狀況並適時地予以協助。
本研究旨在學生於大規模開放線上課程參與課程內容所提供線上課程功能，包括：閱讀線上電子書內容、觀看線上課程影片、線上課程測驗。透過學生連續操作的行為所產生之記錄（log），經資料蒐集、處理、定義後，將其行為簡化為編碼(encode)。例如使用者一連串的學習活動（learning activity）資訊，其內容包括：使用者名稱、Session、操作狀態及發生時間…等。
透過一連串的分析方法，可否根據學生在線上學習平台的操作，將學習活動序列化（sequential），再透過不同的序列模式探勘方法（Sequential Pattern Mining）找出學生共有的學習子序列（learning sub-sequences）？探勘出學生學習子序列後，是否能再經由這些學習子序列找出與學習成效(learning performance)相關性及具有代表性的學習模式（learning pattern）？藉由探索性因素分析（Exploratory Factor Analysis, EFA）的主要成份分析（Principal Component Analysis, PCA）找出學習模式的主要成分為哪些？隨後使用多重因素分析（Multiple Factor Analysis, MFA）觀察不同學習成效的學生其學習成就上的差異為何？
經過研究結果證實，所採用的序列模式探勘方法，確實能夠說明學生的共同學習行為對學習成效具正、負面的影響。並根據其該課程學生在線上課程學習序列模式(Types of learning sequences)，定義出四個主要學習策略(Learning strategies)。

然而，這些結果在將來是否可透過學生在學習線上課程的過程裡，觀察學生的學習模式如果趨向於某種不良的學習策略時，透過線上課程系統的學習輔助設計，給予學生在學習上的建議，已提升學生的學習成效，則是未來可繼續深入探討的方向。

關鍵字：MOOCs（Massive Open Online Courses）、行為序列（activity sequence）、主成份分析（Principal Component Analysis）、探索性因素分析（Exploratory Factor Analysis）、多重因素分析（Multiple Factor Analysis, MFA）

With the arising of technology and the extensive utilization of Internet, online open education resources are also emerging. Massive Open Online Course (MOOC) is the latest remote education indicator recently. The features of MOOC are open access and its scalability. In early 2014, the Ministry of Education of Taiwan developed the MOOCs program. By the open source provided by Coursera, Udacity, and Open edX, various colleges and universities developed their own open online courses, aiming at creating a learning environment for students free from the restriction on the time, space and location. In addition, based on the learning data from the extensive participation of the students, it is intended to provide more assistance to the students according to the learning situation.
This study is directed to the students' participation of the online course functions provided by the large scale MOOC courses, including reading online e-book content, watching online course videos, and the online course tests. Through the records (logs) generated by the students' continuous operation, after the data is collected, processed, and defined, the behavior is simplified into an encoding. For example, the information of a series learning activities by a user includes: user name, Session, operation status, time of occurrence, etc.
Through a series of analysis methods, is it possible to make the learning activities sequential and then to find a common learning sub-sequences between the students according to different sequential patter mining methods (such as, Sequential Patter Mining, etc.)? After mining the students learning sub-sequences, is it possible to find the correlation between the learning performance and the typical learning pattern based on the learning sub-sequences? What are the main components of the learning pattern found by Principal Component Analysis (PCA) of Exploratory Factor Analysis (EFA)? Afterward, what are the learning performance differences by observing the student groups under different performances using the Multiple Factor Analysis (MFA)?
According to the result of the present research, the adopted Sequence Pattern Detections method shows that the students' common learning behavior has a positive/negative effects on learning performances. According to the types of learning sequences of the students on the online courses, four major learning strategies are defined.
For future study, further research may be conducted on how the results of the present solution may aid to provide suggestions to the students by the auxiliary learning system of the online course system in order to improve the learning performance when it is observed that the learning pattern of the students tends to a more improper learning strategy during the online course learning.

Keywords: MOOCs (Massive Open Online Courses), activity sequence, Principal Component Analysis, Exploratory Factor Analysis, Multiple Factor Analysis

一、 緒論 1
1.1 研究背景 1
1.2 研究動機 1
1.3 研究目的 2
二、文獻探討 4
2.1 流程探勘 4
2.2 學習模式偵測流程 5
2.3 使用者行為模式偵測流程 6
2.4 滯後序列分析（Lag Sequential Analysis, LSA） 7
2.5 序列模式探勘（Sequential Pattern Mining） 8
2.5.1 情節探勘（Episode mining - Sliding window） 9
2.5.2 PrefixSpan algorithm 9
三、研究內容與方法 11
3.1 課程描述 11
3.2 資料集描述 11
3.2.1 資料集離群值處理 13
3.3 LMS線上學習流程簡介 14
3.3.1 資料前處理-連續學習活動（Session）之定義 15
3.4 學習模式偵測流程 16
3.4.1 資料前處理階段 16
3.4.1.1 Log產生流程 16
3.4.1.2 LMS之Log各欄位定義 17
3.4.1.3 LMS之Log中各事件狀態（event type）之定義 19
3.4.1.4 線上學習活動種類定義 19
3.4.1.5 學習序列編碼 21
3.4.2 學習序列模式偵測階段 22
3.4.2.1 滯後序列分析 24
3.4.2.2 序列模式探勘-情節探勘 27
3.4.2.3 序列模式探勘-前綴偵測演算法 30
3.4.3 學習子序列選擇條件 35
3.4.4 定義學習子序列的種類 36
3.5 小結 39
四、研究結果與討論 40
4.1 學習序列探勘主要採取的方法 40
4.1.1 學習序列模式探勘方法各別之說明 40
4.1.2 學習序列模式探勘方法之比較 41
4.1.3 學習序列模式探勘方法之選擇 42
4.2 學習子序列與學習成效的相關性 42
4.2.1 全班學習子序列對應種類與說明 43
4.3 說明學生的主要學習模式 48
4.3.1 探索性因素分析之說明 48
4.3.2 探索性因素分析之結果 49
4.3.3 學生主要學習模式之說明 52
4.4 影響學習成效的關鍵學習模式 59
4.4.1 多重因素分析之說明 60
4.5 全班學生學習策略之探勘 64
4.5.1 全班學生學習策略探勘之方法 64
4.5.2 學習序列模式之定義 65
4.5.3 學習序列模式之說明及EFA成份之對應 67
4.5.4 定義學習策略之方法並說明 70
4.5.5 學生學習策略之說明 74
五、結論與未來研究 76
參考文獻 79

Weinstein, C. E., Husman, J., & Dierking, D. R. (2000). Self-regulation interventions with a
focus on learning strategies. In Handbook of self-regulation (pp. 727-747). Academic
Press.Agrawal, R., & Srikant, R. (1995, March). Mining sequential patterns. In icde (Vol.
95, pp. 3-14).
Winne, P. H. (2013). Learning strategies, study skills, and self-regulated learning in
postsecondary education. In Higher education: Handbook of theory and research
(pp. 377-403). Springer, Dordrecht.
Srikant, R., & Agrawal, R. (1996, March).Mining sequential patterns: Generalizations and
performance improvements.In International Conference on Extending Database
Technology (pp. 1-17).Springer, Berlin, Heidelberg.
Blikstein, P., Worsley, M., Piech, C., Sahami, M., Cooper, S., & Koller, D. (2014).
Programming pluralism: Using learning analytics to detect patterns in the learning of
computer programming. Journal of the Learning Sciences, 23(4), 561-599.
Jeong, H., Biswas, G., Johnson, J., & Howard, L. (2010, June). Analysis of productive
learning behaviors in a structured inquiry cycle using hidden Markov models. In
Educational Data Mining 2010.
Lust, G., Elen, J., & Clarebout, G. (2013). Regulation of tool-use within a blended course:
Student differences and performance effects. Computers & Education, 60(1), 385-395.
Jovanović, J., Gašević, D., Dawson, S., Pardo, A., & Mirriahi, N. (2017). Learning analytics
to unveil learning strategies in a flipped classroom. The Internet and Higher
Education, 33(4), 74-85.
Perera, D., Kay, J., Koprinska, I., Yacef, K., & Zaïane, O. R. (2008). Clustering and sequential
pattern mining of online collaborative learning data. IEEE Transactions on Knowledge
and Data Engineering, 21(6), 759-772.
Van Der Aalst, W. (2011). Process mining: discovery, conformance and enhancement of
business processes (Vol. 2). Heidelberg: Springer.
Sackett, G. P. (1979). The lag sequential analysis of contingency and cyclicity in behavioral
interaction research. Handbook of infant development, 1, 623-649.
Pei, J., Han, J., Mortazavi-Asl, B., Pinto, H., Chen, Q., Dayal, U., & Hsu, M. C. (2001).
Prefixspan: Mining sequential patterns efficiently by prefix-projected pattern growth.In
Proceedings 17th international conference on data engineering (pp. 215-224). IEEE.
Pei, J., Han, J., Mortazavi-Asl, B., Wang, J., Pinto, H., Chen, Q., ... & Hsu, M. C. (2004).
Mining sequential patterns by pattern-growth: The prefixspan approach. IEEE
Transactions on knowledge and data engineering, 16(11), 1424-1440.
Gabadinho, A., Ritschard, G., Mueller, N. S., & Studer, M. (2011). Analyzing and visualizing
state sequences in R with TraMineR. Journal of Statistical Software, 40(4), 1-37.
Codish, D., Rabin, E., & Ravid, G. (2019). User behavior pattern detection in unstructured
processes–a learning management system case study. Interactive Learning Environments,
1-27.