網路上的衝突會對用戶造成多方面的不良影響，包括情緒上的壓力、心理健康問題和社交互動的減少。現有的管理方式多半依靠用戶檢舉，但這種方法效率低下且反應較慢。如果採用自動偵測系統，則以自然語言處理或關鍵字偵測為主，導致誤判率高，並可能誤刪正常留言。因此，本研究探討了一種更高準確率的辨識方法，並尋找對正常言論影響較小的衝突行為管理方式。
本研究分為兩大部分。第一部分探討圖神經網路結合社會網路分析和自然語言處理來檢測線上評論中的衝突。我們從Reddit的“worldnews”子版收集評論數據，利用自然語言處理進行情感分析以獲取邊的參數，並使用社會網路分析計算中心度指標來獲取節點參數。GNN模型在檢測性能方面顯示出優秀的正確率。最佳配置包含中心度指標和情感分析分數，突顯了利用多種特徵捕捉微妙互動和關係的重要性。該模型有效地識別了衝突言論而不過度標記正常評論。第二部分為實驗驗證。本實驗將參與者分為未處理、遮蔽和警告三組，並使用正面負面情緒量表測量前後情緒變化。結果顯示，與未處理組相比，遮蔽組和警告組顯著減少了負面情緒，而遮蔽組和警告組之間沒有顯著差異。結果表明，這兩種策略在減少負面情緒影響方面同樣有效。
最後，根據本研究模型辨識和實驗的結果，希望能提供一個良好的社群媒體衝突管理方式。通過這些措施，社交媒體平台可以創造一個更良好的網絡環境，增強用戶的心理健康和正面互動體驗，從而提升用戶在社交媒體平台上的整體福祉。

Incivility on the Internet can adversely affect users, causing emotional stress, mental health issues, and reduced social interactions. Current management methods mainly rely on user reporting, which is inefficient and slow. Automatic detection systems, often using natural language processing or keyword detection, have high false positive rates. This study explores a more accurate identification method and seeks ways to manage incivility with minimal impact on normal comments.
This study has two main parts. The first part investigates using Graph Neural Networks (GNN) combined with Social Network Analysis (SNA) and Natural Language Processing (NLP) to detect incivility. We collected comment data from Reddit's "worldnews" subreddit, using NLP to obtain edge parameters and SNA to calculate for node parameters. The GNN model demonstrated excellent detection accuracy. The optimal configuration included centrality metrics and sentiment analysis scores. The second part involved experimental validation. Participants were divided into three groups: unblocked, blocked, and warning. The PANAS (Positive and Negative Affect Schedule) measured emotional changes before and after the intervention. Results showed that blocked and warning groups significantly reduced negative emotions compared to unblocked group, with no significant difference between blocked and warning groups. The results indicate that both strategies are equally effective in reducing negative emotions.
Based on the identification model and experimental results, we hope to provide a better incivility management approach for social media platforms. These measures can create a better online environment, enhance users' mental health and positive interaction experiences, and improve overall well-being on social media platforms.

Abstract 2
摘要 3
List of Contents 4
1. Introduction 6
1.1 Research Background 6
1.2 Research Motivation 7
1.3 Research Question 8
2. Literature Review 9
2.1 Negative Impact of Incivility 9
2.2 GNN, SNA and NLP for Incivility Detection 10
2.3 Experimental Studies on Effects of Warning Labels 12
3. Methodology 15
3.1 Data Collection 18
3.2 Incivility Detection 21
3.3 An Experiment on the Impact of Incivility Warning Messages 25
4. Data Analysis 28
4.1 Incivility Detection 28
4.2 Impact of Incivility Warning Messages 32
5. Discussion 40
5.1 Incivility Detection 40
5.2 Impact of Incivility Warning Messages 42
6. Conclusion 44
6.1 Limitation 46
6.2 Future Work 48
7. Reference 49
8. Appendix 54

Alberto, T. C., Lochter, J. V., & Almeida, T. A. (2015). Post or block? Advances in automatically filtering undesired comments. Journal of Intelligent & Robotic Systems, 80, 245-259.
Anderson, A. A., Brossard, D., Scheufele, D. A., Xenos, M. A., & Ladwig, P. (2014). The “nasty effect:” Online incivility and risk perceptions of emerging technologies. Journal of computer-mediated communication, 19(3), 373-387.
Anderson, A. A., Yeo, S. K., Brossard, D., Scheufele, D. A., & Xenos, M. A. (2018). Toxic talk: How online incivility can undermine perceptions of media. International Journal of Public Opinion Research, 30(1), 156-168.
Antoci, A., Bonelli, L., Paglieri, F., Reggiani, T., & Sabatini, F. (2019). Civility and trust in social media. Journal of Economic Behavior & Organization, 160, 83-99.
Antoci, A., Delfino, A., Paglieri, F., Panebianco, F., & Sabatini, F. (2016). Civility vs. incivility in online social interactions: An evolutionary approach. PloS one, 11(11), e0164286.
Bacile, T. J., Wolter, J. S., Allen, A. M., & Xu, P. (2018). The effects of online incivility and consumer-to-consumer interactional justice on complainants, observers, and service providers during social media service recovery. Journal of Interactive Marketing, 44(1), 60-81.
Băroiu, A. C., & Trăușan-Matu, Ș. (2023). Comparison of deep learning models for automatic detection of sarcasm context on the MUStARD dataset. Electronics, 12(3), 666.
Benslimane, S., Azé, J., Bringay, S., Servajean, M., & Mollevi, C. (2023). A text and GNN based controversy detection method on social media. World Wide Web, 26(2), 799-825.
Bianchi, F., Hills, S. A., Rossini, P., Hovy, D., Tromble, R., & Tintarev, N. (2022). " It's Not Just Hate'': A Multi-Dimensional Perspective on Detecting Harmful Speech Online. arXiv preprint arXiv:2210.15870.
Bormann, M. (2022). Perceptions and evaluations of incivility in public online discussions—Insights from focus groups with different online actors. Frontiers in Political Science, 4, 812145.
Busch, J., Kocheturov, A., Tresp, V., & Seidl, T. (2021, July). NF-GNN: network flow graph neural networks for malware detection and classification. In Proceedings of the 33rd International Conference on Scientific and Statistical Database Management (pp. 121-132).
Chen, Y., & Wang, L. (2022). Misleading political advertising fuels incivility online: A social network analysis of 2020 US presidential election campaign video comments on YouTube. Computers in Human Behavior, 131, 107202.
Davidson, S., Sun, Q., & Wojcieszak, M. (2020, November). Developing a new classifier for automated identification of incivility in social media. In Proceedings of the fourth workshop on online abuse and harms (pp. 95-101).
Ding, R. X., Wang, X., Shang, K., & Herrera, F. (2019). Social network analysis-based conflict relationship investigation and conflict degree-based consensus reaching process for large scale decision making using sparse representation. Information Fusion, 50, 251-272.
Dittmar, H., Halliwell, E., & Stirling, E. (2009). Understanding the impact of thin media models on women's body-focused affect: The roles of thin-ideal internalization and weight-related self-discrepancy activation in experimental exposure effects. Journal of social and clinical psychology, 28(1), 43-72.
Erlanger, S. (2009, December 2). Point, shoot, retouch and label? https://www.nytimes.com/2009/12/03/fashion/03Boyer.html
Gervais, B. T. (2015). Incivility online: Affective and behavioral reactions to uncivil political posts in a web-based experiment. Journal of Information Technology & Politics, 12(2), 167-185.
Ghosh, D., Vajpayee, A., & Muresan, S. (2020). A report on the 2020 sarcasm detection shared task. arXiv preprint arXiv:2005.05814.
Hamilton, W., Ying, Z., & Leskovec, J. (2017). Inductive representation learning on large graphs. Advances in neural information processing systems, 30.
Jin, D., Yu, Z., Jiao, P., Pan, S., He, D., Wu, J., ... & Zhang, W. (2021). A survey of community detection approaches: From statistical modeling to deep learning. IEEE Transactions on Knowledge and Data Engineering, 35(2), 1149-1170.
Kipf, T. N., & Welling, M. (2016). Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907.
Korenčić, D., Baris, I., Fernandez, E., Leuschel, K., & Salido, E. S. (2021, April). To block or not to block: Experiments with machine learning for news comment moderation. In Proceedings of the EACL Hackashop on News Media Content Analysis and Automated Report Generation (pp. 127-133).
Muchnik, L., Aral, S., & Taylor, S. J. (2013). Social influence bias: A randomized experiment. Science, 341(6146), 647-651.
Mueller, C. E., & Winsor, D. L. (2018). Depression, suicide, and giftedness: Disentangling risk factors, protective factors, and implications for optimal growth. Handbook of giftedness in children: psychoeducational theory, research, and best practices, 255-284.
Nandwani, P., & Verma, R. (2021). A review on sentiment analysis and emotion detection from text. Social network analysis and mining, 11(1), 81.
Ng, Y. L., Song, Y., Kwon, K. H., & Huang, Y. (2020). Toward an integrative model for online incivility research: A review and synthesis of empirical studies on the antecedents and consequences of uncivil discussions online. Telematics and informatics, 47, 101323.
Nishi, K. (2024). Negative Impact of Online Political Incivility on Willingness to See Political Comments. arXiv preprint arXiv:2403.08372.
Pareja, A., Domeniconi, G., Chen, J., Ma, T., Suzumura, T., Kanezashi, H., ... & Leiserson, C. (2020, April). Evolvegcn: Evolving graph convolutional networks for dynamic graphs. In Proceedings of the AAAI conference on artificial intelligence (Vol. 34, No. 04, pp. 5363-5370).
Pennycook, G., Bear, A., Collins, E. T., & Rand, D. G. (2020). The implied truth effect: Attaching warnings to a subset of fake news headlines increases perceived accuracy of headlines without warnings. Management science, 66(11), 4944-4957.
Posavac, H. D., Posavac, S. S., & Weigel, R. G. (2001). Reducing the impact of media images on women at risk for body image disturbance: Three targeted interventions. Journal of social and clinical psychology, 20(3), 324-340.
Reporter, G. S. (2020, July 1). Ban airbrushing of models in children’s ads, says MP. The Guardian. https://www.theguardian.com/politics/2009/aug/03/airbrush-models-advertising-jo-swinson?CMP=share_btn_url
Sadeque, F., Rains, S., Shmargad, Y., Kenski, K., Coe, K., & Bethard, S. (2019, June). Incivility detection in online comments. In Proceedings of the eighth joint conference on lexical and computational semantics (* SEM 2019) (pp. 283-291).
Salminen, J., Hopf, M., Chowdhury, S. A., Jung, S. G., Almerekhi, H., & Jansen, B. J. (2020). Developing an online hate classifier for multiple social media platforms. Human-centric Computing and Information Sciences, 10, 1-34.
Slater, A., Tiggemann, M., Firth, B., & Hawkins, K. (2012). Reality check: An experimental investigation of the addition of warning labels to fashion magazine images on women's mood and body dissatisfaction. Journal of Social and Clinical Psychology, 31(2), 105-122.
Stoll, A., Ziegele, M., & Quiring, O. (2020). Detecting impoliteness and incivility in online discussions: Classification approaches for German user comments. Computational Communication Research, 2(1), 109-134.
Su, X., Xue, S., Liu, F., Wu, J., Yang, J., Zhou, C., ... & Philip, S. Y. (2022). A comprehensive survey on community detection with deep learning. IEEE Transactions on Neural Networks and Learning Systems.
Sun, Q., Wojcieszak, M., & Davidson, S. (2021). Over-time trends in incivility on social media: evidence from political, non-political, and mixed sub-reddits over eleven years. Frontiers in Political Science, 3, 741605.
Veličković, P., Cucurull, G., Casanova, A., Romero, A., Lio, P., & Bengio, Y. (2017). Graph attention networks. arXiv preprint arXiv:1710.10903.
Walsh, S. (2024, July 4). The Top 10 Social Media Sites & Platforms. Search Engine Journal. https://www.searchenginejournal.com/social-media/social-media-platforms/
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: the PANAS scales. Journal of personality and social psychology, 54(6), 1063.
Weinstein, E. (2017). Adolescents' differential responses to social media browsing: Exploring causes and consequences for intervention. Computers in Human Behavior, 76, 396-405.
Wu, Z., Pan, S., Chen, F., Long, G., Zhang, C., & Philip, S. Y. (2020). A comprehensive survey on graph neural networks. IEEE transactions on neural networks and learning systems, 32(1), 4-24.