Massive open online courses (MOOC) 是個新興的學習模式，它變得越來越熱門因為學習中時間和地點上的限制被去除了。然而新問題經常伴隨著新方法而來，MOOC學習者在學習過程中常遇上一些困難，例如認知過載的問題。知識地圖的概念是處理認知過載問題的有效方法之一，它使用了知識可視化的技術提供有效的學習策略。然而認知過載的問題仍然存在，因為學習者不一定知道課程素材與知識地圖中的哪些重要知識概念相關。舉例來說，對於初學者，判斷一次考試的題目想要考核什麼知識概念是相當困難的。在這個研究中，我們使用了不同的課程素材如課本與講義來訓練本系統，之後本系統能夠自動連結其他課程素材到知識地圖中的知識概念，如考試題目與課程影片。我們透過幾種文字探勘的技術分析訓練文本來建立關係模型，並以此判斷隨機測試文本與知識的關聯程度。實驗證明了我們提出的自動分類方法比人工分類快速而且也足夠準確。當有良好的訓練集，系統的準確度(Accuracy)可達90%而且精確度(Precision)可達80%。這樣的特色對於學習者很有幫助，因為他們可以簡單快速的了解任何課程教材在講述什麼重要的知識概念。

Massive open online courses (MOOC) is an uprising learning method which is more and more popular because it eliminates the limitation of places and time for learning. However, new problems often come with new methods. Some difficulties have been observed on the learners of MOOC such as cognitive overload. The use of knowledge maps is one of the powerful technique to deal with cognitive overload. Knowledge maps provide effective learning strategies by visualizing the knowledge to a certain structure. However, the problems of cognitive overload are still there because the learners may not know the relationship between certain course materials and the important concepts in the knowledge map. For instance, it is quite hard for new learners to pinpoint what concept is being asked in an exam question. In this study, we use different kinds of course materials such as textbooks and slides to train the system first. And then the system is able to automatically map other course materials such as exam questions and course videos into the concepts in knowledge map. We perform several text mining techniques over the training corpora to build some models, and then determine the relevance between the concepts in the knowledge maps and random testing corpora. Our experiments proved the auto mapping mechanism is faster than manual mapping, and it is also accurate. With a proper training set, the accuracy will be higher than 90%, and the precision will be higher than 80%. This feature is really helpful for learners since they can identify what knowledge concepts are mentioned in any course materials easily.

Abstract ........................................................................................................................ i
中文摘要....................................................................................................................... ii
Contents .................................................................................................................. iii
Chapter 1 Introduction .............................................................................................. 1
1.1 What is MOOC? ............................................................................................. 1
1.2 About Knowledge Maps …............................................................................. 2
1.3 What is Course Material Mapping? ................................................................ 3
1.3.1 Problem Statement ...................................................................................... 3
1.3.2 Definition of Concept Mapping .................................................................. 4
1.4 AI Tutor big picture .................................................…………….………….. 7
Chapter 2 Related Work ............................................................................................ 9
2.1 MOOC ............................................................................................................ 9
2.1.1 Coursera ............................................................................................ 10
2.1.2 edX......................................................................................................10
2.1.3 Udacity .............................................................................................. 11
2.2 Knowledge Maps .......................................................................................... 12
2.3 Student Types Analysis ................................................................................. 12
2.4 Student Sentiment Analysis .......................................................................... 13
2.5 Student Interests Analysis ............................................................................ 13
2.6 Course Forum Classification ..................................................................... 14
2.7 Text Mining In Other Fields ...................................................................... 14
Chapter 3 System Architecture ............................................................................ 15
3.1 System Overview ...................................................................................... 15
3.2 System Components in Detail ................................................................... 16
3.3 System Workflow ...................................................................................... 17
3.3.1 Interpretation of the Workflow ....................................................... 18
3.3.2 Workflow in Detail ......................................................................... 19
3.3.2.1 Training Phase ..................................................................... 19
3.3.2.2 Mapping Phase ....................................................................... 19
3.3.3 Example .......................................................................................... 20
3.3.3.1 Training Phase ..................................................................... 20
3.3.3.2 Mapping Phase .................................................................... 21
Chapter 4 System Implementation ...................................................................... 22
4.1 Core Processing System ............................................................................ 22
4.1.1 Data Set Preprocessor ..................................................................... 22
4.1.1.1 Text Trimming ..................................................................... 22
4.1.1.2 Text Segmentation ............................................................... 23
4.1.1.3 Stop Words Deletion ............................................................ 24
4.1.1.4 Text Normalization .............................................................. 25
4.1.2 Corpora Translation Module .......................................................... 25
4.1.3 String Comparison Module ............................................................ 26
4.1.4 Topic Modeling Module ................................................................. 27
4.1.4.1 Concept ................................................................................ 27
4.1.4.2 Implementation .................................................................... 29
4.1.4.2.1 Proposed Topic Modeling Module .............................. 29
4.1.4.2.2 Semi-Mapping ............................................................. 30
4.1.4.2.3 Experts Modification System ...................................... 31
4.1.5 Corpora Structure Analyzer ............................................................... 31
4.1.6 Automatic Training Set Generator ..................................................... 32
4.1.7 Word2vec Module ............................................................................. 33
4.1.7.1 What are word vectors? .......................................................... 34
4.1.7.2 What is word2vec? ................................................................. 34
4.1.7.3 Algorithms of word2vec ......................................................... 35
4.1.7.3.1 What is Neuron Network? ........................................... 35
4.1.7.3.2 Continuous Bag-of-Words (CBOW) ........................... 36
4.1.7.3.3 Skip-gram .................................................................... 37
4.1.7.3.4 Hierarchical Softmax and Negative Sampling ............ 38
4.1.7.4 Implementation ....................................................................... 38
4.1.7.4.1 Proposed Word2vec module ........................................ 38
4.1.7.4.2 Semi-Mapping ............................................................. 39
4.1.8 Map Module ...................................................................................... 40
4.1.8.1 Check Corpora Structure ........................................................ 40
4.1.8.2 Semi-Result Structure ............................................................ 41
4.1.8.3 Weights of Semi-Mapping Result .......................................... 41
4.1.8.3.1 Automatic Fitting ........................................................ 41
4.1.8.3.2 Adjustment by Experience .......................................... 42
4.1.8.4 Determine the Mapping Result .............................................. 42
4.1.8.5 Result Handling ...................................................................... 43
4.1.8.5.1 Course Video Mapping .....................………………... 43
4.1.8.5.2 Batch Inputs Mapping ................................................. 43
4.1.8.5.3 Real-time Corpora Mapping ........................................ 43
4.1.9 Accuracy Calculator .......................................................................... 44
4.1.10 Discuss ............................................................................................ 45
4.1.10.1 String Comparison ................................................................ 46
4.1.10.2 Word2vec .............................................................................. 46
4.1.10.3 Topic Model .......................................................................... 46
4.1.10.4 Other Algorithms .................................................................. 47
4.2 Data Server ................................................................................................... 47
4.2.1 Database ............................................................................................ 47
4.2.2 Server File System ............................................................................ 48
4.2.3 Fetch Label Module .......................................................................... 48
4.3 Web Server ................................................................................................... 49
4.4 User Interface Module .................................................................................. 50
4.4.1 Real-time Corpora Mapping .............................................................. 52
4.4.1.1 Experts Modification Module ................................................ 53
4.4.2 Batch Mapping and Video Subtitle Mapping .................................... 54
Chapter 5 Experiment .............................................................................................. 55
5.1 Course Choosing .......................................................................................... 55
5.1.1 Training Materials ............................................................................. 55
5.1.2 Courses Difference ............................................................................ 56
5.2 Principles of Economics ............................................................................... 56
5.2.1 Training Set and Testing Set .............................................................. 56
5.2.2 Result ................................................................................................. 57
5.3 Introduction to Computer Network .............................................................. 57
5.3.1 Training Set and Testing Set .............................................................. 57
5.3.2 Result ................................................................................................. 58
Chapter 6 ................................................................................................................... 60
6.1 Key Reasons that Affect Result .................................................................... 60
6.1.1 Labeled Training Corpora ................................................................. 60
6.1.2 Knowledge Maps ............................................................................... 60
6.1.2.1 Well-Structured Knowledge Maps ......................................... 61
6.1.2.2 Not Well-Structured Knowledge Maps .................................. 61
6.1.3 Corpora with Images.......................................................................... 62
6.2 How to improve the precision? .................................................................... 63
Chapter 7 Conclusion and Future Work ................................................................ 65
Bibliography ............................................................................................................. 67

[1] A Ng and D. Koller, "Coursera," Retrieved July 12, 2018, from the World Wide Web: https://www.coursera.org/
[2] M. I. of Technology and H. University, "Open edX," Retrieved July 12, 2018, from the World Wide Web: https://www.edx.org/
[3] M.S. Sebastian Thrun, David Stavens, "Udacity," Retrieved July 12, 2018, from the World Wide Web: https://www.udacity.com/
[4] Minhong Wang, Jun Peng, Bo Cheng, Hance Zhou and Jie Liu, "Knowledge visualization for self-regulated learning." Educational Technology & Society, Vol. 14, No. 3, pp. 28-42, 2011.
[5] José A. Ruipérez-Valiente, Pedro J. Muñoz-Merino, Carlos Delgado Kloos, Katja Niemann, Maren Scheffel, and Martin Wolpers, "Analyzing the impact of using optional activities in self-regulated learning." IEEE Transactions on Learning Technologies, vol. 9, no. 3, pp. 231-243, 2016.
[6] Ayse Saliha Sunar, Su White, Nor Aniza Abdullah, and Hugh C. Davis, "How learners' interactions sustain engagement: a mooc case study." IEEE Transactions on Learning Technologies, vol. 10, no. 4, 2017.
[7] Charles D. Holley and Donald F. Dansereau, "Spatial learning strategies: Techniques, applications, and related issues." Academic Press, 2014.
[8] National Tsing Hua University, "Sharecourse," Retrieved July 12, 2018, from the World Wide Web: http://www.sharecourse.net/sharecourse/
[9] Reviews.com, "The Best MOOC Platforms of 2018," Retrieved July 12, 2018, from the World Wide Web: https://www.reviews.com/mooc-platforms/
[10] T.O. University, "FutureLearn," Retrieved July 12, 2018, from the World Wide Web: https://www.futurelearn.com/
[11] Jonas Liepmann and Hannes Klöpper, "iversity," Retrieved July 12, 2018, from the World Wide Web: https://iversity.org/
[12] IBM, "Cognitive Class," Retrieved July 12, 2018, from the World Wide Web: https://cognitiveclass.ai/
[13] Sal Khan, "Khan Academy," Retrieved July 12, 2018, from the World Wide Web: https://www.khanacademy.org/
[14] F. X. ZHOU, "Junyi Academy," Retrieved July 12, 2018, from the World Wide Web: https://www.junyiacademy.org/
[15] Zhuoxuan Jiang, Yan Zhang, and Xiaoming Li, "MOOCon: A Framework for Semi-supervied Concept Extraction from MOOC Content." Database Systems for Advanced Applications, 2017.
[16] "Tf-idf," Retrieved July 12, 2018, from the World Wide Web: https://zh.wikipedia.org/wiki/Tf-idf
[17] "Conditional random field," Retrieved July 12, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Conditional_random_field
[18] Nen-Fu Huang, Chia-An Lee, Yi-Wei Huang, Po-Wen Ou, How-Hsuan Hsu, So-Chen Chen, and Jian-Wei Tzeng, "On the Automatic Construction of Knowledge-Map from handouts for MOOC Courses." Smart Innovation, Systems and Technologies, vol 81. 2017.
[19] Seungwhan Moon, Saloni Potdar, and Lara Martin, "Identifying Student Leaders from MOOC Discussion Forums through Language Influence." Proceedings of the Empirical Methods in Natural Language Processing Workshop, pp. 15-20. 2014.
[20] Wen, Miaomiao, Yang, Diyi and Rosé C.P., "Sentiment Analysis in MOOC Discussion Forums: What does it tell us?" Proceedings of Educational Data Mining, pp. 130-137, 2014.
[21] Arti Ramesh , Dan Goldwasser , Bert Huang , Hal Daume ́ Iii , and Lise Getoor, "Understanding MOOC Discussion Forums using Seeded LDA." Proceedings of the 9th ACL Workshop on Innovative Use of NLP for Building Educational Applications, 2014.
[22] Ling Wang, Gongliang Hu and Tiehua Zhou, "Semantic Analysis of Learners’Emotional Tendencies on Online MOOC Education." Sustainability, vol. 10, 2018.
[23] X. Wei, H. Lin, L. Yang and Y. Yu, "A Convolution-LSTM-Based Deep Neural Network for Cross-Domain MOOC Forum Post Mapping" Information, 2017.
[24] "Transfer learning," Retrieved July 12, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Transfer_learning
[25] "Convolutional neural network," Retrieved July 12, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Convolutional_neural_network
[26] "Long short-term memory," Retrieved July 12, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Long_short-term_memory
[27] Xiang Feng and Longhui Qiu, "The Construction of Sentiment Lexicon in Educational Field Based on Word2vec." Annual Convention of the Association for Educational Communications and Technology, vol 1, 2017.
[28] "Word2vec," Retrieved July 12, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Word2vec
[29] Feng Yu and Dequan Zheng, "Education Data Mining: How to Mine Interactive Text in MOOCs using Natural Language Process." 12th International Conference on Computer Science and Education (ICCSE), 2017.
[30] "k-nearest neighbors algorithm," Retrieved July 12, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/K-nearest_neighbors_algorithm
[31] "Support vector machine," Retrieved July 12, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Support_vector_machine
[32] Xian Peng, Sanya Liu, Zhi Liu, Wenbin Gan and Jianwen Sun, "MINING LEARNERS' TOPIC INTERESTS IN COURSE REVIEWS BASED ON LIKE-LDA MODEL." International Journal of Innovative Computing, vol. 12, no. 6, 2016.
[33] Rel Guzman Apaza, Elizabeth Vera Cervantes, Laura Cruz Quispe, and Jose ́ Ochoa Luna, "Online courses recommendation based on LDA." 1st Symposium on Information Management and Big Data, 2014.
[34] Thushari Atapattu and Katrina Falkner, "Framework for Topic Generation and Labeling from MOOC Discussions." Proceedings of the Third ACM Conference on Learning, pp. 201-204, 2016.
[35] Aysu Ezen-Can, Kristy Elizabeth Boyer, Shaun Kellogg and Sherry Booth, "Unsupervised Modeling for Understanding MOOC Discussion Forums: A Learning Analytics Approach." Proceedings of the Fifth International Conference on Learning Analytics and Knowledge, pp. 146-150, 2015.
[36] "k-medoids," Retrieved July 12, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/K-medoids
[37] K.W. Lim, C. Chen and W.L. Buntine, "Twitter-Network Topic Model: A Full Bayesian Treatment for Social Network and Text Modeling." Advances in Neural Information Processing Systems: Topic Models Workshop, pp.1-5, 2013.
[38] "Gaussian process," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Gaussian_process
[39] Vitomir Kovanović, Srećko Joksimović, Dragan Gašević, George Siemens and Marek Hatala, "What public media reveals about MOOCs: A systematic analysis of news reports." British Journal of Educational Technology, 2014.
[40] "Node.js," Retrieved July 13, 2018, from the World Wide Web: https://nodejs.org/en/
[41] "python," Retrieved July 13, 2018, from the World Wide Web: https://www.python.org/
[42] "TensorFlow," Retrieved July 13, 2018, from the World Wide Web: https://www.tensorflow.org/
[43] "Flask," Retrieved July 13, 2018, from the World Wide Web: http://flask.pocoo.org/
[44] "Jade," Retrieved July 13, 2018, from the World Wide Web: http://jade-lang.com/
[45] "MySQL," Retrieved July 13, 2018, from the World Wide Web: https://www.mysql.com/
[46] "Representational state transfer," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Representational_state_transfer
[47] "PDFtoText," Retrieved July 13, 2018, from the World Wide Web: https://pdftotext.com/
[48] National Tsing Hua University, "Introduction to computer networks," Retrieved July 13, 2018, from the World Wide Web: http://www.sharecourse.net/sharecourse/course/view/courseInfo/1406
[49] "scikit-learn," Retrieved July 13, 2018, from the World Wide Web: http://scikit-learn.org/stable/
[50] "Jieba," Retrieved July 13, 2018, from the World Wide Web: https://github.com/fxsjy/jieba
[51] "NLTK," Retrieved July 13, 2018, from the World Wide Web: https://www.nltk.org/
[52] "python-inflection," Retrieved July 13, 2018, from the World Wide Web: https://github.com/jrbl/python-infection/blob/master/infection.py
[53] "Latent Dirichlet allocation," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Latent_Dirichlet_allocation
[54] David M. Blei, Andrew Y. Ng, Michael I. Jordan, "Latent Dirichlet allocation." Journal of Machine Learning Research, pp.993-1022, 2003
[55] "One-hot," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/One-hot
[56] Tomas Mikolov, Kai Chen, Greg Corrado and Jeffrey Dean, "Efficient Estimation of Word Representations in Vector Space." ICLR Workshop Papers, 2013.
[57] "Artificial neural network," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Artificial_neural_network
[58] "Learning Word Vectors from Sherlock Holmes," Retrieved July 13, 2018, from the World Wide Web: https://pat-coady.github.io/word2vec/
[59] "Cosine similarity," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Cosine_similarity
[60] "Confusion matrix," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Confusion_matrix
[61] "Generative adversarial network," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Generative_adversarial_network
[62] "Decision tree," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Decision_tree
[63] "k-means clustering," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/K-means_clustering
[64] "Natural language processing," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Natural_language_processing
[65] "Recurrent neural network," Retrieved July 13, 2018, from the World Wide Web: https://en.wikipedia.org/wiki/Recurrent_neural_network
[66] National Tsing Hua University, "Principles of Economics (I)," Retrieved July 13, 2018, from the World Wide Web: http://www.sharecourse.net/sharecourse/course/view/courseInfo/1577