溝通誤解的主因是，純文字在解讀時，讀者會因爲不同的背景文化，而可能產生與作者想表達的不同的理解。更精確地說，溝通誤解的發生並不是作者的本意，而是作者在無意間，因為寫作架構所傳達的寫作風格，讓讀者產生歧義感受所致。因此，本研究的主要目標在於自寫作結構中，辨別出可能導致溝通誤解的寫作風格。作為一個相對較新的議題，目前並沒有可以直接使用的溝通誤解數據集。此外，現有對寫作結構進行分析的研究方法，很容易受到單個作者寫作風格的影響，導致分類器在訓練時產生偏差。針對這兩個挑戰，我們首先設計了數據集構建的規則。之後，我們採用基於圖論建構文字圖型想法，改良並提出「均衡分佈圖型（Low Disperison Patterns）」方法，用以解構與描述在溝通誤解的情況中普遍使用的書寫結構。根據實驗結果，本研究所提出的均衡分佈圖型，改善了先前因分佈特徵不平衡引起的不穩定擬合問題，並得到了令人滿意的誤傳預測結果。

Miscommunications happen while readers' perception and the writer's intention are inconsistent when reading plain text. More precisely, miscommunications are not caused by writer's intention but unintentionally triggered by the writing styles expressed through the writing structure. Therefore, the main goal of this research is to identify the writing structure that may lead to miscommunications. However, as a relatively new topic, no dataset is available. Besides, the previous studies for analyzing the writing structure are easily affected by the writing style of a single author, which causes the classifier to bias during training. For the two challenges, we firstly design the rule for the dataset construction. Afterward, we adapt the graph-based pattern approach to propose the low dispersion (LD) patterns method for decomposing and describe the common writing structure. According to the experiment results, the proposed LD patterns improve the unstable fitting problem caused by the unbalanced distributed features and get satisfied results for miscommunication prediction.

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2 Related Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

3 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.1 Miscommunication Dataset Construction . . . . . . . . . . . . . . . . . . . . 8
3.1.1 Defining Conversation . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.1.2 Intention Capturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.1.3 Miscommunication Determination . . . . . . . . . . . . . . . . . . . 12
3.2 Miscommunication Patterns Extraction . . . . . . . . . . . . . . . . . . . . . 12
3.2.1 Graph Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2.2 Pattern Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.3 Writing Style Representations . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.3.1 Pattern Score Representation . . . . . . . . . . . . . . . . . . . . . . . 20
3.3.2 Widening Pattern Representation . . . . . . . . . . . . . . . . . . . . 22
3.3.3 Low Dispersion Pattern Representation . . . . . . . . . . . . . . . . . 23

4 Experiment and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.1 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.1.1 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.1.2 Data Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
4.2 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.2.1 Word Selection: Connector Word And Subject . . . . . . . . . . . . 29
4.2.2 Miscommunication Prediction . . . . . . . . . . . . . . . . . . . . . . 31
4.3 Dataset Limitation Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . 36

5 Conclusion and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

[1] Sweta Agrawal and Amit Awekar. Deep learning for detecting cyberbullying across multiple social media platforms. In European Conference on Information Retrieval, pages 141–153. Springer, 2018.

[2] Carlos Argueta, Elvis Saravia, and Yi-Shin Chen. Unsupervised graph-based patterns extraction for emotion classification. In Proceedings of the 2015 IEEE/ACM Interna- tional Conference on Advances in Social Networks Analysis and Mining 2015, pages 336–341. ACM, 2015.

[3] Edward Loper Bird, Steven and Ewan Klein. Natural language processing with python. In The Oxford handbook of internet studies. 2009.

[4] David M Blei, Andrew Y Ng, and Michael I Jordan. Latent dirichlet allocation. Journal of machine Learning research, 3(Jan):993–1022, 2003.

[5] No´e Cecillon, Vincent Labatut, Richard Dufour, and Georges Linar`es. Abusive lan- guage detection in online conversations by combining content-and graph-based fea- tures. Frontiers in Big Data, 2:8, 2019.

[6] Yung-Chun Chang, Cen-Chieh Chen, Yu-Lun Hsieh, Chien Chin Chen, and Wen-Lian Hsu. Linguistic template extraction for recognizing reader-emotion and emotional resonance writing assistance. In Proceedings of the 53rd Annual Meeting of the Asso- ciation for Computational Linguistics and the 7th International Joint Conference on Natural Language Processing (Volume 2: Short Papers), volume 2, pages 775–780, 2015.

[7] Justin Cheng, Cristian Danescu-Niculescu-Mizil, and Jure Leskovec. Antisocial be- havior in online discussion communities. In Ninth International AAAI Conference on Web and Social Media, 2015.

[8] Cristian Danescu-Niculescu-Mizil, Moritz Sudhof, Dan Jurafsky, Jure Leskovec, and Christopher Potts. A computational approach to politeness with application to social factors. arXiv preprint arXiv:1306.6078, 2013.

[9] Harsh Dani, Jundong Li, and Huan Liu. Sentiment informed cyberbullying detection in social media. pages 52–67, 2017.

[10] Fabio Del Vigna, Andrea Cimino, Felice Dell’Orletta, Marinella Petrocchi, and Maurizio Tesconi. Hate me, hate me not: Hate speech detection on facebook. 01 2017.

[11] Eduard Dragut, Fang Fang, Prasad Sistla, Clement Yu, and Weiyi Meng. Stop word and related problems in web interface integration. Proceedings of the VLDB Endow- ment, 2(1):349–360, 2009.

[12] Nicole B Ellison and Danah M Boyd. Sociality through social network sites. In The Oxford handbook of internet studies. 2013.

[13] Mai ElSherief, Vivek Kulkarni, Dana Nguyen, William Yang Wang, and Elizabeth Belding. Hate lingo: A target-based linguistic analysis of hate speech in social media. In Twelfth International AAAI Conference on Web and Social Media, 2018.

[14] Esteban, Ortiz-Ospina. The rise of social media, 2019. data retrieved from Our World in Data, https://ourworldindata.org/rise-of-social-media.

[15] Tin Kam Ho. Bootstrapping text recognition from stop words. In Proceedings. Four- teenth International Conference on Pattern Recognition (Cat. No. 98EX170), volume 1, pages 605–609. IEEE, 1998.

[16] Pamela Jordan and Richmond H Thomason. Refining the categories of miscommuni- cation. In In Proceedings of AAAI Workshop on Detecting, Repairing, and Preventing Human-machine Miscommunication, CA. Citeseer, 1996.

[17] Aniket Kittur, Ed Chi, and Bongwon Suh. What’s in wikipedia? mapping topics and
conflict using socially annotated category structure. 04 2009.

[18] Shervin Malmasi and Marcos Zampieri. Detecting hate speech in social media. In Proceedings of the International Conference Recent Advances in Natural Language Processing, RANLP 2017, pages 467–472, Varna, Bulgaria, September 2017. IN- COMA Ltd.

[19] Binny Mathew, Ritam Dutt, Pawan Goyal, and Animesh Mukherjee. Spread of hate speech in online social media. In Proceedings of the 10th ACM Conference on Web Science, pages 173–182. ACM, 2019.

[20] Vinita Nahar, Xue Li, and Chaoyi Pang. An effective approach for cyberbullying detection. Communications in Information Science and Management Engineering, 3(5):238, 2013.

[21] Vinita Nahar, Sayan Unankard, Xue Li, and Chaoyi Pang. Sentiment analysis for effective detection of cyberbullying. In Asia-Pacific Web Conference, pages 767– 774. Springer, 2012.

[22] Chikashi Nobata, Joel Tetreault, Achint Thomas, Yashar Mehdad, and Yi Chang. Abusive language detection in online user content. In Proceedings of the 25th in- ternational conference on world wide web, pages 145–153. International World Wide Web Conferences Steering Committee, 2016.

[23] Jennifer Rihn and Jonathan James Oliver. Detection of abusive user accounts in social networks, October 30 2018. US Patent App. 10/116,614.

[24] Elvis Saravia, Hsien-Chi Toby Liu, Yen-Hao Huang, Junlin Wu, and Yi-Shin Chen. Carer: Contextualized affect representations for emotion recognition. In Proceedings ofthe2018ConferenceonEmpiricalMethodsinNaturalLanguageProcessing, pages 3687–3697, 2018.

[25] Yi-jie Tang and Hsin-Hsi Chen. Mining sentiment words from microblogs for pre-
dicting writer-reader emotion transition. In LREC, pages 1226–1229, 2012.

[26] Zeerak Waseem, Thomas Davidson, Dana Warmsley, and Ingmar Weber. Understanding abuse: A typology of abusive language detection subtasks. In Proceedings of the
First Workshop on Abusive Language Online, pages 78–84, Vancouver, BC, Canada, August 2017. Association for Computational Linguistics.

[27] Naomi Yamashita and Toru Ishida. Automatic prediction of misconceptions in multi- lingual computer-mediated communication. In Proceedings of the 11th international conference on Intelligent user interfaces, pages 62–69. ACM, 2006.

[28] ANK Zaman, Pascal Matsakis, and Charles Brown. Evaluation of stop word lists in text retrieval using latent semantic indexing. In 2011 Sixth International Conference on Digital Information Management, pages 133–136. IEEE, 2011.

[29] Marcos Zampieri, Shervin Malmasi, Preslav Nakov, Sara Rosenthal, Noura Farra, and Ritesh Kumar. Predicting the type and target of offensive posts in social media. arXiv preprint arXiv:1902.09666, 2019.

[30] Justine Zhang, Jonathan Chang, Cristian Danescu-Niculescu-Mizil, Lucas Dixon, Yiqing Hua, Dario Taraborelli, and Nithum Thain. Conversations gone awry: Detect- ing early signs of conversational failure. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1350– 1361, Melbourne, Australia, July 2018. Association for Computational Linguistics.