喜歡就點讚？探討社群媒體中符號與文字情感之影響力

Like to Click? The Influence of Emotions in Emoji and Text on Social Media

10.6846/TKU.2017.00903

曾筱喬

表情符號 ； 文章品質 ； 文章熱門度 ； Emoji ； Quality of Post ； Popularity of Post

淡江大學企業管理學系碩士班學位論文

2017年

碩士

張瑋倫

繁體中文

2011年iOS行動裝置系統支援表情符號(Emoji)功能，便帶來溝通媒介革新的開始，在行動裝置日益普及之背景下，順水推舟使Emoji使用率逐年成長，經Emoji Consumer Science Team於2015調查使用Emoji之原因得知，其中以“可以幫助對方更清楚的了解訊息”為最主要原因。Tauch and Kanjo（2016）認為短訊傳遞情感不如於面對面交流方式，因此，許多人利用表情符號表達自己的情緒，表情符號相較於本文更富有情感更貼近使用者感覺和心情。而在Facebook具有相當豐富之情緒符號數據之平台，Like符號於2009年開放至今，且於2015年又加入其它五種表情符號。因此，本研究欲建構符號與文字情感為基礎之情感分析模式，及探討符號與文字中情感之文章熱門程度對於社會網絡之影響力，發展以情感為基礎之文章熱門程度之模式，將表情符號與文字進行量化，因此，採用SentiStrength軟體與LIWC軟體，轉換Facebook上六種表情符號與文字為數字，並將表情符號視為權重之概念，而文字情感為輔助之角色，以減少情感因素之誤差，並將情感因素建構在Chen, Hheng, He and Jiang (2012) 所提文章品質之概念，認為具有影響力的文章是需要以品質的文章為基礎，好品質的文章是尤分享文章次數及回文的數量為衡量依據。 本研究沿用此概念為文章品質熱門度，發展出個人文章熱門度與主題文章熱門度，根據所建立模式，而本研究的研究對象為2016美國總統大選候選人希拉蕊與川普將所收集之發文，分為四種類別選舉相關議題、希拉蕊與川普討論對手相關議題、希拉蕊與川普討論個人相關議題、其它相關議題，經本研究所提出主題評論模式比較結果得知，川普在討論對手相關議題、討論個人相關議題、其它相關議題分數皆高於希拉蕊，而希拉蕊在選舉相關議題則高於川普，本研究認為此結果與希拉蕊為政治家背景有密且關係，在具有娛樂性質之社交平台中，追蹤政治家之動機，則會較偏向為支持態度或是對於政治議題較關心之使用者，因此，在談論選舉之話題較引起使用者之共鳴，當僅考慮文章品質熱門度與本研究文章熱門度與主題文章熱門度模式比較時，文章品質熱門度能呈現出文章討論度高低，但無法了解在這高討論度是由憤怒或喜愛等情感所匯聚而成，但本研究所發展之模式，因具有情感因素，透由情感因素可放大或縮小文章品質熱門度，便可了解文章之情感導向。在情感對社會網絡影響力之關係中，利用轉發為社會網絡之特徵探討表情符號與轉發之關係，根據情符號與分享次數之相關係數結果得知，以Like及Love與分享次數相關性最高且基於轉發為社會網絡影響力之特徵與因素之一，同時也代表情感因素與社會網絡影響力關係為正相關，本研究亦進一步將社會網絡範圍擴大至Google關鍵字搜尋趨勢圖上，檢視情感因素、分享與時事之關係，比較情感因素與分享對於時事解釋力強度如何，結果得知，總表情符號數相較於分享次數與Google趨勢圖之關係較為有強烈關係，其中又以選舉議題與Google趨勢圖關係皆較有中度強烈正相關，表示在社會媒體中情感回饋相較於分享與時事正相關更高。

In 2011, iOS devices supported the emojisin the text which pushed the increment of usage. According to the report of Consumer Science Team in 2015, emojis can help users clearly understand the message. Tauch and Kanjo (2016) considered short message (text) may not deliver the emotions compared to face-to-face communication. That is, emojis may present relevant emotions and close to real feelings. In 2015, Facebook launched new functions that allowed users to use five more emoticons except “like”. This research considers the emotions in the text and emojis to investigate the influence of a post on social media. We build a sentiment-based model to measure the influence of a post on social media. The sentiment in text and emoticon will be quantified by using SentiStrength and LIWC in the proposed model. The concept was to take into account emoticons as the weight to adjust the quality of post based on the research of Chen, Hheng, He and Jiang (2012). This study developed a model that combines quality of post and sentiment of text and emoticons simultaneously. Our model has two different types: individual and joint topic for comparison. We collected data on Facebook of 2016 US presidential campaign between September and November 8th with two major candidates: Hillary Clinton and Donald Trump. The collected data was separated into four categories: election, opponent, individual, and others. The results showed Donald Trump has higher scores of opponent, individual, and others than Hillary Clinton. Hillary Clinton has higher score of election than Donald Trump. We infer that the background of Hillary Clinton may be the reason to attract followers to support the posts of election issue. We discovered quality of post only can reveal the popularity of a post with unknown positive or negative emotions. Our model (individual and joint topic) can not only reveal the popularity of the post but also present the direction of emotion. Furthermore, the outcomes showed emoticons has positive relations with number of share; particularly, Like and Love have highest impact on the relations. In other words, our model proves the positive influence of emoticons on a post in social media. In addition, we examine the relations between total number of emoji and number of share. The outcomes showed total number of emoji has higher positive relations than number of share on the curve of Google trend. The category of election has highest impact on the relations (the curve of Google trend) and shows emojis have positive relations on social media.

1. 1. Alshenqeeti, H. (2016). Are Emojis Creating a New or Old Visual Language for New Generations? A Socio-semiotic Study. Advances in Language and Literary Studies, 7(6), 56-69.
   連結：
2. 5. Chen, W., Cheng, S., He, X., & Jiang, F. (2012, November). Influencerank: An efficient social influence measurement for millions of users in microblog. In Cloud and Green Computing (CGC), 2012 Second International Conference on (pp. 563-570). IEEE.
   連結：
3. 7. Francalanci, Chiara, and Ajaz Hussain. "Influence-based Twitter browsing with NavigTweet." Information Systems 64 (2017): 119-131.
   連結：
4. 9. Kavanagh, B. (2016). Emoticons as a medium for channeling politeness within American and Japanese online blogging communities. Language & Communication, 48, 53-65.
   連結：
5. 12. Lahuerta-Otero, E., & Cordero-Gutiérrez, R. (2016). Looking for the perfect tweet. The use of data mining techniques to find influencers on twitter. Computers in Human Behavior, 64, 575-583.
   連結：
6. 14. Munger, T., & Zhao, J. (2015, August). Identifying influential users in on-line support forums using topical expertise and social network analysis. In 2015 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM) (pp. 721-728). IEEE.
   連結：
7. 16. Miller, H., Thebault-Spieker, J., Chang, S., Johnson, I., Terveen, L., & Hecht, B. (2016). “Blissfully happy” or “ready to fight”: Varying Interpretations of Emoji. ICWSM’16.
   連結：
8. 17. Moschini, I. (2016). The" Face with Tears of Joy" Emoji. A Socio-Semiotic and Multimodal Insight into a Japan-America Mash-Up. HERMES-Journal of Language and Communication in Business, (55), 11-25.
   連結：
9. 22. Ravi, K., & Ravi, V. (2015). A survey on opinion mining and sentiment analysis: Tasks, approaches and applications. Knowledge-Based Systems, 89, 14-46.
   連結：
10. 25. Settanni, M., & Marengo, D. (2015). Sharing feelings online: studying emotional well-being via automated text analysis of Facebook posts. Frontiers in psychology, 6.
    連結：
11. 27. Skiba, D. J. (2016). Face with Tears of Joy Is Word of the Year: Are Emoji a Sign of Things to Come in Health Care?. Nursing education perspectives, 37(1), 56-57.
    連結：
12. 29. Stieglitz, S., & Dang-Xuan, L. (2012, January). Political communication and influence through microblogging--An empirical analysis of sentiment in Twitter messages and retweet behavior. In System Science (HICSS), 2012 45th Hawaii International Conference on (pp. 3500-3509). IEEE.
    連結：
13. 33. Wang, H., Lei, K., & Xu, K. (2015, June). Profiling the followers of the most influential and verified users on Sina Weibo. In 2015 IEEE International Conference on Communications (ICC) (pp. 1158-1163). IEEE.
    連結：
14. 34. Wu, F., Huang, Y., Song, Y., & Liu, S. (2016). Towards building a high-quality microblog-specific Chinese sentiment lexicon. Decision Support Systems.
    連結：
15. 38. Zhao, W. X., Liu, J., He, Y., Lin, C. Y., & Wen, J. R. (2014, August). A computational approach to measuring the correlation between expertise and social media influence for celebrities on microblogs. In Advances in Social Networks Analysis and Mining (ASONAM), 2014 IEEE/ACM International Conference on (pp. 460-463). IEEE.
    連結：
16. 6. Internet Live Stats
    連結：
17. 12. Oxford Dictionaries（2015）Word of the Year 2015 is…
    連結：
18. 英文部分
19. 2. Alp, Z. Z., & Öğüdücü, Ş. G. (2016, November). Influential user detection on Twitter: Analyzing effect of focus rate. In Advances in Social Networks Analysis and Mining (ASONAM), 2016 IEEE/ACM International Conference on (pp. 1321-1328). IEEE.
20. 3. Bakshy, E., Hofman, J. M., Mason, W. A., & Watts, D. J. (2011, February). Everyone's an influencer: quantifying influence on twitter. In Proceedings of the fourth ACM international conference on Web search and data mining (pp. 65-74). ACM.
21. 4. Bhargav, M., & Bhargav, A. (2014, July). Mining relationships from text in social networking sites. In Control, Instrumentation, Communication and Computational Technologies (ICCICCT), 2014 International Conference on (pp. 31-35). IEEE.
22. 6. Cui, A., Zhang, M., Liu, Y., & Ma, S. (2011, December). Emotion tokens: Bridging the gap among multilingual twitter sentiment analysis. In Asia Information Retrieval Symposium (pp. 238-249). Springer Berlin Heidelberg.
23. 8. Hill, S., Benton, A., Ungar, L., Macskassy, S., Chung, A., & Holmes, J. H. (2016). A Cluster-based Method for Isolating Influence on Twitter.
24. 10. Khatua, A., Khatua, A., Ghosh, K., & Chaki, N. (2015, January). Can# Twitter_Trends Predict Election Results? Evidence from 2014 Indian General Election. In System Sciences (HICSS), 2015 48th Hawaii International Conference on (pp. 1676-1685). IEEE.
25. 11. Komrsková, Z. (2015). The Use of Emoticons in Polite Phrases of Greeting and Thanks. International Journal of Social, Behavioral, Educational, Economic, Business and Industrial Engineering, 9(4), 1309-1312.
26. 13. Liu, B. (2012). Sentiment analysis and opinion mining. Synthesis lectures on human language technologies, 5(1), 1-167.
27. 15. Mittal, S., Goel, A., & Jain, R. (2016, October). Sentiment analysis of E-commerce and social networking sites. In Computing for Sustainable Global Development (INDIACom), 2016 3rd International Conference on (pp. 2300-2305). IEEE.
28. 18. Naveed, N., Gottron, T., Kunegis, J., & Alhadi, A. C. (2011, June). Bad news travel fast: A content-based analysis of interestingness on twitter. In Proceedings of the 3rd International Web Science Conference (p. 8). ACM.
29. 19. Niu, T., Zhu, S., Pang, L., & El Saddik, A. (2016, January). Sentiment analysis on multi-view social data. In International Conference on Multimedia Modeling (pp. 15-27). Springer International Publishing.
30. 20. Novak, P. K., Smailović, J., Sluban, B., & Mozetič, I. (2015). Sentiment of emojis. PloS one, 10(12), e0144296.
31. 21. Pavalanathan, U., & Eisenstein, J. (2015). Emoticons vs. emojis on Twitter: A causal inference approach. arXiv preprint arXiv:1510.08480.
32. 23. Romero, D. M., Galuba, W., Asur, S., & Huberman, B. A. (2011, September). Influence and passivity in social media. In Joint European Conference on Machine Learning and Knowledge Discovery in Databases (pp. 18-33). Springer Berlin Heidelberg.
33. 24. Roy, S. D., & Zeng, W. (2014, July). Influence of social media on performance of movies. In Multimedia and Expo Workshops (ICMEW), 2014 IEEE International Conference on (pp. 1-6). IEEE.
34. 26. Severyn, A., Moschitti, A., Uryupina, O., Plank, B., & Filippova, K. (2016). Multi-lingual opinion mining on youtube. Information Processing & Management, 52(1), 46-60.
35. 28. Smailović, J., Kranjc, J., Grčar, M., Žnidaršič, M., & Mozetič, I. (2015, October). Monitoring the Twitter sentiment during the Bulgarian elections. In Data Science and Advanced Analytics (DSAA), 2015. 36678 2015. IEEE International Conference on (pp. 1-10). IEEE.
36. 30. Soranaka, K., & Matsushita, M. (2012, November). Relationship between emotional words and emoticons in tweets. In 2012 Conference on Technologies and Applications of Artificial Intelligence (pp. 262-265). IEEE.
37. 31. Tauch, C., & Kanjo, E. (2016, September). The roles of emojis in mobile phone notifications. In Proceedings of the 2016 ACM International Joint Conference on Pervasive and Ubiquitous Computing: Adjunct (pp. 1560-1565). ACM.
38. 32. Wang, H., & Castanon, J. A. (2015, October). Sentiment expression via emoticons on social media. In Big Data (Big Data), 2015 IEEE International Conference on (pp. 2404-2408). IEEE.
39. 35. Ye, S., & Wu, S. F. (2010, October). Measuring message propagation and social influence on Twitter. com. In International Conference on Social Informatics (pp. 216-231). Springer Berlin Heidelberg.
40. 36. Yeole, A. V., Chavan, P. V., & Nikose, M. C. (2015, March). Opinion mining for emotions determination. In Innovations in Information, Embedded and Communication Systems (ICIIECS), 2015 International Conference on (pp. 1-5). IEEE.
41. 37. Yuan, Z., & Purver, M. (2015). Predicting emotion labels for chinese microblog texts. In Advances in Social Media Analysis (pp. 129-149). Springer International Publishing.
42. 39. Ali, A., & Chan, E. C. (2016). The Key to Coaching. Learning, Application and Practice. Lulu. com.
43. 40. Fiske, J., & 張錦華. (1995). 傳播符號學理論. 台北: 遠流, 62-63.
44. 網站部分
45. 1. http://www.business2community.com/social-media/social-media-growth-statistics-01545217#qXFAbHMgZeAMJHKH.97
46. 2. Jonny Rosen（ 2016，May）Social Media Growth Statistics. Business2Community
47. 3. https://engineering.instagram.com/emojineering-part-1-machine-learning-for-emoji-trendsmachine-learning-for-emoji-trends-7f5f9cb979ad#.5tk4k1b7a
48. 4. Instagram Engineering（2015, May）Emojineering Part 1: Machine Learning for Emoji Trends.
49. 5. http://www.internetlivestats.com/internet-users/#sources
50. 7. http://marketingland.com/facebook-usage-accounts-1-5-minutes-spent-mobile-171561
51. 8. Greg Sterling ( 2016,April ) Nearly 80 percent of social media time now spent on mobile devices. Marketing Land
52. 9. https://www.appboy.com/blog/emojis-used-in-777-more-campaigns/
53. 10. Jesse Tao（2016,March）EMOJIS ARE NOW USED IN 777% MORE CAMPAIGNS THAN LAST YEAR 
54. 11. http://blog.oxforddictionaries.com/2015/11/word-of-the-year-2015-emoji/