隨著線上平台的使用頻率增加，確實的溝通也愈加重要。但是由於溝通媒介的限制，有時文字上的溝通會造成誤解並引起衝突。在本論文中，我們嘗試事先辨識出可能引起錯誤表達的文字訊息並且修正它們。傳統的分類技術難以解釋訊息中會產生誤解的部份，除此之外，缺乏標註的錯誤表達資料也使得這項研究更具挑戰性。我們提出一個利用群眾智慧的做法以辨識在社群平台Reddit上造成錯誤表達的使用者貼文，並建構一個框架能夠辨識導致衝突的文字模式以預測上述情形。

As usage of online platforms increases, effective conversations become of great importance. Unfortunately, given the limited medium of communication, written text is at times misunderstood and generate conflict. In this thesis we attempt to identify such miscommunication before they occur and rephrase messages before it is too late. Traditional classification methods do not provide interpretability which is necessary to identify what parts of the message generate misunderstanding. Furthermore, the lack of annotated data about miscommunication provides an additional challenge. We propose a method to identify posts on the social media platform Reddit that cause miscommunication and conflict using wisdom of the crowd, as well as a framework that allows the prediction of these cases by identifying patterns that lead to conflict.

Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2 RelatedWork . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3 DataCollection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4 FrameworkDefinition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.1 Pattern Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.1.1 Weights Construction . . . . . . . . . . . . . . . . . . . . . . . . . 14
4.1.2 Weights Aggregation . . . . . . . . . . . . . . . . . . . . . . . . . 15
4.1.3 Graph Creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.1.4 Graph Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.1.5 Pattern Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.1.6 Pattern Enrichment . . . . . . . . . . . . . . . . . . . . . . . . . . 19
4.2 Pattern Ranking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.3 Conflict Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
5 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
5.2 Data Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5.3 Experimental Setup and Results . . . . . . . . . . . . . . . . . . . . . . . 26
5.4 Patterns Extracted . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
A ThresholdValues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

[1] Anqi Cui, Min Zhang, Yiqun Liu, and Shaoping Ma. Emotion tokens: Bridging the gap among multilingual twitter sentiment analysis. In Asia information retrieval symposium, pages 238–249. Springer, 2011.
[2] Fabio Del Vigna12, Andrea Cimino23, Felice Dell’Orletta, Marinella Petrocchi, and Maurizio Tesconi. Hate me, hate me not: Hate speech detection on facebook. 2017.
[3] Justin Kruger, Nicholas Epley, Jason Parker, and Zhi-Wen Ng. Egocentrism over e-mail: Can we communicate as well as we think? Journal of personality and social psychology, 89(6):925, 2005.
[4] Shervin Malmasi and Marcos Zampieri. Detecting hate speech in social media. arXiv preprint arXiv:1712.06427, 2017.
[5] Tomas Mikolov, Kai Chen, Greg Corrado, and Jeffrey Dean. Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781, 2013.
[6] George Miller. WordNet: An electronic lexical database. MIT press, 1998.
[7] Vinita Nahar, Xue Li, and Chaoyi Pang. An effective approach for cyberbullying detection. Communications in Information Science and Management Engineering, 3(5):238, 2013.
[8] Vinita Nahar, Sayan Unankard, Xue Li, and Chaoyi Pang. Sentiment analysis for effective detection of cyber bullying. In Asia-Pacific Web Conference, pages 767–774. Springer, 2012.
[9] Chikashi Nobata, Joel Tetreault, Achint Thomas, Yashar Mehdad, and Yi Chang. Abusive language detection in online user content. In Proceedings of the 25th in international conference on world wide web, pages 145–153. International World Wide Web Conferences Steering Committee, 2016.
[10] Jeffrey Pennington, Richard Socher, and Christopher Manning. Glove: Global vectors for word representation. In Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP), pages 1532–1543, 2014.
[11] Ellen Riloff and Janyce Wiebe. Learning extraction patterns for subjective expressions. In Proceedings of the 2003 conference on Empirical methods in natural language processing, pages 105–112. Association for Computational Linguistics, 2003.
[12] Elvis Saravia, Carlos Argueta, and Yi-Shin Chen. Unsupervised graph-based pattern extraction for multilingual emotion classification. Social Network Analysis and Mining, 6(1):92, 2016.
[13] Elvis Saravia, Hsien-Chi Toby Liu, and Yi-Shin Chen. Deepemo: Learning and enriching pattern-based emotion representations. arXiv preprint arXiv:1804.08847, 2018.
[14] Piet Van Mieghem. Human psychology of common appraisal: The reddit score. IEEE Transactions on Multimedia, 13(6):1404–1406, 2011.
[15] William Warner and Julia Hirschberg. Detecting hate speech on the world wide web. In Proceedings of the Second Workshop on Language in Social Media, pages 19–26. Association for Computational Linguistics, 2012.
[16] Zeerak Waseem, Thomas Davidson, Dana Warmsley, and Ingmar Weber. Understanding abuse: a typology of abusive language detection subtasks. arXiv preprint arXiv:1705.09899, 2017.
[17] Ellery Wulczyn, Nithum Thain, and Lucas Dixon. Ex machina: Personal attacks seen at scale. In Proceedings of the 26th International Conference on World Wide Web, pages 1391–1399. International World Wide Web Conferences Steering Committee, 2017.
[18] Ainur Yessenalina, Yisong Yue, and Claire Cardie. Multi-level structured models for document-level sentiment classification. In Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing, pages 1046–1056. Association for Computational Linguistics, 2010.
[19] Dawei Yin, Zhenzhen Xue, Liangjie Hong, Brian D Davison, April Kontostathis, and Lynne Edwards. Detection of harassment on web 2.0. Proceedings of the Content Analysis in the WEB, 2:1–7, 2009.
[20] Justine Zhang, Jonathan P Chang, Cristian Danescu-Niculescu-Mizil, Lucas Dixon, Yiqing Hua, Nithum Thain, and Dario Taraborelli. Conversations gone awry: Detecting early signs of conversational failure. arXiv preprint arXiv:1805.05345, 2018.