回聲室效應是指一人的想法逐漸得到強化的過程，而我們日常使用的社群網站恰好 就為我們帶來了這樣的一個環境，而這樣的一個環境也將使我們的視野逐漸限制在 和我們相似的使用者之間。因此，為了使我們了解在我們所處的環境中回聲室效應 的程度，本研究以Facebook使用者的貼文和其留言為基礎，藉由了解使用者以及多 個留言的立場狀況，來推測該使用者的回聲室效應程度。更進一步地說，我們偵測 使用者每一篇貼文的echoing數值，依照不同事件觀察其數值變化，並將同一使用 者各貼文數值統整成一回聲室效應指標以作為本研究之最終結果。在立場偵測的部 分，我們提出了以「情緒」以及「目標立場」為兩大Features，來幫助我們偵測一 貼文以及一留言的立場狀況。另一方面，在Feature Extraction的動作中，我們亦提 出了以Graph-based Pattern為基礎，讓我們只以一種形態的Pattern來提取Features。 整個實驗中，主要以Facebook粉絲專頁為主要資料來源，藉由了解粉絲專頁中的回 聲室效應來作為判斷實驗成果的基礎。

Echo chamber is a metaphorical description of a situation in which beliefs are amplified inside a closed network, and social media platforms just provide a good environment for the situation. However, if echo chamber goes serious in social media platforms, our judgement is restricted from different opinions. Hence, in this work, we focus on finding the degree of the echo chamber for a facebook user, which is based on detecting the echoing interaction between a post and its related comments. On the other hands, we detect the echoing score for each post, and observe tendency of the echo chamber along with events. In the experiment, we utilize two features, including emotion and target stance, to detect echoing situation. Besides, in the feature extraction, we propose the graph-based pattern to extract the feature. In the data collection, we focus on Facebook fan page as our main data source to evaluate the results.

1 Introduction .................................... 1
2 RelatedWork.................................... 6
2.1 EchoChamber................................. 6
2.2 StanceDetection ............................... 7
2.3 EmotionAnalysisonSocialMedia...................... 9
3 DataCollection................................... 11
3.1 Overview ................................... 11
3.2 TarEmoPatternConstructionDataCollection . . . . . . . . . . . . . . . . 12
3.3 TargetStanceDataCollection......................... 12
3.4 EmotionDataCollection ........................... 13
3.5 EchoChamberDataCollection........................ 14
4 Methodology.................................... 16
4.1 Overview ................................... 16
4.2 TarEmoPatternExtraction .......................... 18
4.2.1 GraphConstruction.......................... 19
4.2.2 GraphAggregation .......................... 19
4.2.3 TokenCategorization......................... 19
4.2.4 PatternCandidate........................... 21
4.2.5 PatternEnrichment .......................... 22
4.2.6 Pre-trainedWordEmbeddings .................... 23
4.2.7 TarEmo Pattern Construction by Mingling Emotion-relevant Characteristic................................ 23
4.3 PatternVectorization ............................. 24
4.3.1 pf-itf (pattern frequency-inverse target stance frequency) . . . . . . 25
4.3.2 pf-ief (pattern frequency-inverse emotion frequency) . . . . . . . . 26
4.4 FeaturesExtractionfromTarEmoPattern................... 28
4.4.1 Target Stance Feature Extraction Neural Network . . . . . . . . . . 28
4.4.2 EmotionFeatureExtractionNeuralNetwork. . . . . . . . . . . . . 29
4.5 Echo Chamber Indexing Based on Two Features from TarEmo Pattern . . . 30 4.5.1 ECHOModel............................. 30
4.5.2 EchoChamberIndexCalculation................... 32
5 Experiment&Results............................... 34
5.1 Overview ................................... 34
5.1.1 Evaluationmethod .......................... 34
5.1.2 Experiments.............................. 35
5.2 Echoing/NotEchoingDetection ....................... 36
5.2.1 Experiment Results in Echoing/Not Echoing Detection . . . . . . . 36
5.2.2 Echoing Tendency Comparison in News Fan Page . . . . . . . . . . 40
5.3 StanceDetection ............................... 43
5.3.1 Overview ............................... 43
5.3.2 ExperimentResultsofAgreementDetection . . . . . . . . . . . . . 44
6 CONCLUSIONSANDFUTUREWORK .................... 47
7 Appendixes..................................... 49
7.1 EchoChamberIndexforSpecificEvent ................... 50
References ....................................... 52

[1] Carlos Argueta, Fernando H Calderon, and Yi-Shin Chen. Multilingual emotion clas- sifier using unsupervised pattern extraction from microblog data. Intelligent Data Analysis, 20(6):1477–1502, 2016.
[2] Katherine A. Heller Charles Blundell, Jeff Beck. Modelling reciprocating relation- ships with hawkes processes. In Proceedings of Advances in Neural Information Processing Systems (NIPS 2012), 2012.
[3] Padhraic Smyth Christopher Dubois, Carter T. Butts. Stochastic blockmodeling of relational event dynamics. In Proceedings of the Sixteenth International Conference on Artificial Intelligence and Statistics, pages 238–246, 2013.
[4] Dmitry Davidov, Oren Tsur, and Ari Rappoport. Enhanced sentiment learning using twitter hashtags and smileys. In Proceedings of the 23rd international conference on computational linguistics: posters, pages 241–249. Association for Computational Linguistics, 2010.
[5] Lucchi A. De Luca V. Severyn A. Muller S. Cieliebak M. Hofmann T. Deriu, J. and M. Jaggi. Leveraging large amounts of weakly supervised data for multi-language sentiment classification. In International World Wide Web Conference Steering Com- mittee, pp, pages 1045–1052, 2017.
[6] Adam Arvidsson Elanor Colleoni, Alessandro Rozza. Echo chamber or public sphere? predicting political orientation and measuring political homophily in twit- ter using big data. Journal of Communication ISSN, 2014.
[7] Rui Fan, Jichang Zhao, Yan Chen, and Ke Xu. Anger is more influential than joy: Sentiment correlation in weibo. PloS one, 9(10):e110184, 2014.
[8] Hanna Wallach Katherine Heller Fangjian Guo, Charles Blundell. The bayesian echo chamber: Modeling social influence via linguistic accommodation. In Proceedings of the 18th International Conference on Artificial Intelligence and Statistics (AISTATS), pages 315–323, 2015.
[9] S. M. Gerrish and D. M. Blei. A language based approach to measuring scholarly impact. In Proceedings of the 26th International Conference on Machine Learning.
[10]AlecGo,RichaBhayani,andLeiHuang.Twittersentimentclassificationusingdistant supervision. CS224N Project Report, Stanford, 1(2009):12, 2009.
[11] K. Zha H. Zhou and L. Song. Learning triggering kernels for multi-dimensional hawkes processes. In Proceedings of the 30th International Conference on Machine Learning, pages 1301–1309, 2013.
[12] Kazi Saidul Hasan and Vincent Ng. Stance classification of ideological debates: Data, models, features, and constraints. In Proceedings of the Sixth International Joint Conference on Natural Language Processing, 2013.
[13] XiaHu,JiliangTang,HuijiGao,andHuanLiu.Unsupervisedsentimentanalysiswith emotional signals. In Proceedings of the 22nd international conference on World Wide Web, pages 607–618. ACM, 2013.
[14] S. W. Linderman and R. P. Adams. Discovering latent network structure in point process data. International Conference on Machine Learning(ICML), 2014.
[15] Andrew Lipsman, Graham Mudd, Mike Rich, and Sean Bruich. The power of “like”. Journal of Advertising research, 52(1):40–52, 2012.
[16] Bing Liu and Lei Zhang. A survey of opinion mining and sentiment analysis. Mining Text Data, Springer, pages 415–463, 2012.
[17] Robert Abbott Ricky Grant Marilyn A. Walker, Pranav Anand. Stance classification using dialogic properties of persuasion. In Proceedings of the 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, page 592–596, 2012a.
[18] Saif M. Mohammad. Sentiment analysis: Detecting valence, emotions, and other affectual states from text. Emotion Measurement, pages 201–237, 2016b.
[19] Alvaro Ortigosa, Jose ́ M Mart ́ın, and Rosa M Carro. Sentiment analysis in facebook and its application to e-learning. Computers in Human Behavior, 31:527–541, 2014.
[20] Saif M. Mohammad Parinaz Sobhani and Svetlana Kiritchenko. Stance and senti- ment in tweets. Special Section of the ACM Transactions on Internet Technology on Argumentation in Social Media, 17(3), 2017.
[21] P. O. Perry and P. J. Wolfe. Point process modelling for directed interaction networks. Journal of the Royal Statistical Society: Series B (Methodology), 75(5)(821–849), 2013.
[22] Rob Abbott Jean E. Fox Tree Robeson Bowmani Michael Minor Pranav Anand, Mar- ilyn Walker. Cats rule and dogs drool!: Classifying stance in online debate. In Pro- ceedings of the 2nd Workshop on Computational Approaches to Subjectivity and Sen- timent Analysis, 2011.
[23] Francisco Rangel and Paolo Rosso. On the identification of emotions and authors’ gender in facebook comments on the basis of their writing style. ESSEM 2013”, CEUR Workshop Proceedings, 1096:34–46, 2013.
[24] Parinaz Sobhani Saif M. Mohammad and Svetlana Kiritchenko. Detecting stance in tweets and analyzing its interaction with sentiment. In Proceedings of the Joint Conference on Lexical and Computational Semantics (*Sem), August 2016.
[25] Elvis Saravia, Hsien-Chi Toby Liu, Yen-Hao Huang, Junlin Wu, and Yi-Shin Chen. Carer: Contextualized affect representations for emotion recognition. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 3687–3697, 2018.
[26] A. Simma and M. I. Jordan. Modeling events with cascades of poisson processes. In Proceedings of the Twenty-Sixth Conference on Uncertainty in Artificial Intelligence, pages 546–555, 2010.
[27] Swapna Somasundaranand and Janyce Wiebe. Recognizing stances in ideological on- line debates. NAACL HLT 2010 Workshop on Computational Approaches to Analysis and Generation of Emotion in Text, pages 116–124, June 2010.
[28] Duyu Tang, Bing Qin, Ting Liu, and Zhenghua Li. Learning sentence representa- tion for emotion classification on microblogs. In Natural Language Processing and Chinese Computing, pages 212–223. Springer, 2013.
[29] Zoubin Ghahramani Tomoharu Iwata, Amar Shah. Discovering latent influence in online social activities via shared cascade poisson processes. In Proceedings of the 19th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pages 266–274, 2013.
[30] Wenbo Wang, Lu Chen, Krishnaprasad Thirunarayan, and Amit P Sheth. Harnessing twitter” big data” for automatic emotion identification. In Privacy, Security, Risk and Trust (PASSAT), 2012 International Conference on and 2012 International Confer- nece on Social Computing (SocialCom), pages 587–592. IEEE, 2012.
[31] Yaowei Wang, Yanghui Rao, Xueying Zhan, Huijun Chen, Maoquan Luo, and Jian Yin. Sentiment and emotion classification over noisy labels. Knowledge-Based Sys- tems, 111:207–216, 2016.
[32] Jichang Zhao, Li Dong, Junjie Wu, and Ke Xu. Moodlens: an emoticon-based senti- ment analysis system for chinese tweets. In Proceedings of the 18th ACM SIGKDD international conference on Knowledge discovery and data mining, pages 1528–1531. ACM, 2012.