疫情期間，大眾媒體是民眾獲得疫病相關資訊的主要消息來源，眾多傳播媒介中，又以新聞產製之疫情相關資訊在此段時間受較高的大眾依賴。與之相應的，在傳染病擴散階段，新聞媒體產製的內容會以議題設定、框架效果等方法型塑之風險意識、影響民眾的行為。旅運模式的變化是都市中重要的非藥物干預之預防措施（NPI），然而過往討論疫情期間由媒體內容改變旅運行為的研究，在傳播媒介上大多限於社交媒體，顯少關注新聞造成之影響，與疫情期間臺灣民眾實際的媒體使用習慣有異，亦尚無文獻實際將媒體製播的疫情資訊的文本分析結果，用作預測旅運工具運量增減的影響因素，而僅探討公眾對交通政策的輿情。綜上，本研究旨在探討共存階段COVID-19疫情及其新聞對臺北捷運運量的影響特性及程度，並弭平過往研究缺口。
在新聞文本的分析上，本文使用自注意力機制（Self-Attention Mechanism）的BERT對新聞進行主題及語意分類。因其預訓練的機制及龐大的運算基礎，BERT是近些年被自然語言處理的應用型研究廣泛使用的大型語言模型。為了取得疫情期間的新聞文本資料，本研究利用瀑布流網頁的網路爬蟲方法爬取研究時間段內近16萬8千則數位新聞，並以標記的新聞類別、文本情緒訓練資料集微調BERT，實現蒐集新聞的分類任務。
校估疫情趨勢與新聞框架效果對臺北捷運系統運量影響時，因疫情趨勢與新聞框架造成的影響並非立即性、且影響不會僅止單一天，必須將時間滯後、延續兩特性納入考量。同時，新聞框架效果本身受疫情趨勢影響，但也會影響臺北捷運運量的中介特性，使得校估具時間特性的時間序列模型無法應用於本研究架構中。為此，本文選用偏最小平方的結構方程模型（PLS-SEM）驗證研究假設之結構路徑。該方法在演算上借鏡主成分分析，將多變數經矩陣轉換降維的特性，並進一步利用迭代運算，實現特徵降維與優化對依變項的預測能力的同步處理，且亦可運用在結構方程模型的校估上。
研究結果顯示，疫情趨勢會對捷運運量產生直接、及經新聞框架中介影響之間接效果；疫情前、後期，組成疫情趨勢及新聞框架效果兩潛在變數之外顯變項的落後期數有明顯差異：前期受落後6至9期之疫情趨勢、落後2至4期之新聞框架效果；後期則受落後1至2期之疫情趨勢、落後1期之新聞框架效果。經本研究驗證的研究期間分段拆分點，與其對應的傳染病擴散階段，可作為理解大眾旅運行為變化之依據。此外，本研究亦驗證新聞文本情緒對捷運運量變動量的影響之調節效果。本研究之分析架構和結果，可協助制訂疫情期間的防疫推廣政策及規範，或供大眾運輸設施排程參考。

During the pandemic, the public has increased their reliance on the news, and the risk perception shaped by media outlets influences preventive behaviors. Previous research discussing the influence of media framing on travel behavior largely focused on social media and rarely addressed the impact of news, differing from media consumption habits of Taiwanese people during the pandemic. This study aims to explore the temporal characteristics and extent of the impact of the coexisting phase of the COVID-19 pandemic and its news on Taipei Metro ridership. In terms of news text classification, this paper uses a text mining model with a Self-Attention Mechanism to categorize and semantically analyze 167,886 online news articles.

Given the lag and continuity of time, the assessment of the impact of pandemic trends and news framing on ridership is verified using the Partial Least Squares Structural Equation Modeling (PLS-SEM) to test the structural paths of the research hypothesis. The results indicate that pandemic trends have direct effects on metro ridership and indirect effects mediated through news framing. In the pre-pandemic and post-pandemic periods, the lag periods of observed variables that constitute pandemic trends and news framing effects differ significantly. Verification of the study period segmentation points can help understand the basis for changes in public travel behavior. Moreover, this study also verifies the moderating effect of the sentiment in news texts on the fluctuations in metro ridership. The analytical framework and results of this study can assist in formulating pandemic prevention promotional policies and guidelines or serve as a reference for public transportation scheduling.

[1] Baruch Fischhoff, “MANAGING RISK PERCEPTIONS,” Issues Sci. Technol. , FALL 1985, Vol. 2, No. 1 (FALL 1985), pp. 83- 96, vol. 21, no. 4, p. 162, 1985.
[2] Q.Ye, K.Ozbay, F.Zuo, andX.Chen, “Impact of Social Media on Travel Behaviors during the COVID-19 Pandemic: Evidence from New York City,” Transp. Res. Rec., vol. 2677, no. 4, pp. 219–238, 2023, doi: 10.1177/03611981211033857.
[3] W.Zha, Q.Ye, J.Li, andK.Ozbay, “A social media Data-Driven analysis for transport policy response to the COVID-19 pandemic outbreak in Wuhan, China,” Transp. Res. Part A Policy Pract., vol. 172, no. March, p. 103669, 2023, doi: 10.1016/j.tra.2023.103669.
[4] X.Chen, Z.Wang, andX.Di, “Sentiment Analysis on Multimodal Transportation during the COVID-19 Using Social Media Data,” Inf., vol. 14, no. 2, pp. 1–12, 2023, doi: 10.3390/info14020113.
[5] M. A.Habib andM. A. H.Anik, “Impacts of COVID-19 on Transport Modes and Mobility Behavior: Analysis of Public Discourse in Twitter,” Transp. Res. Rec. J. Transp. Res. Board, p. 036119812110299, 2021, doi: 10.1177/03611981211029926.
[6] Q.Ye, K.Ozbay, F.Zuo, andX.Chen, “Impact of Social Media on Travel Behaviors during the COVID-19 Pandemic: Evidence from New York City,” Transp. Res. Rec. J. Transp. Res. Board, p. 036119812110338, 2021, doi: 10.1177/03611981211033857.
[7] S.Das andS.Sarkar, “News Media Mining to Explore Speed-Crash-Traffic Association During COVID-19,” Transp. Res. Rec. J. Transp. Res. Board, p. 036119812211212, 2022, doi: 10.1177/03611981221121261.
[8] R.Güner, İ.Hasanoğlu, andF.Aktaş, “Covid-19: Prevention and control measures in community,” Turkish J. Med. Sci., vol. 50, no. SI-1, pp. 571–577, 2020, doi: 10.3906/sag-2004-146.
[9] J. A.Quesada, A.López-Pineda, V. F.Gil-Guillén, J. M.Arriero-Marín, F.Gutiérrez, andC.Carratala-Munuera, “Incubation period of COVID-19: A systematic review and meta-analysis,” Rev. Clínica Española (English Ed., vol. 221, no. 2, pp. 109–117, 2021, doi: https://doi.org/10.1016/j.rceng.2020.08.002.
[10] E.Pieri, “Media Framing and the Threat of Global Pandemics: The Ebola Crisis in UK Media and Policy Response,” Sociol. Res. Online, vol. 24, no. 1, pp. 73–92, 2019, doi: 10.1177/1360780418811966.
[11] A.Casero-Ripollés, “Impact of COVID-19 on the Media System. Communicative and Democratic Consequences of News Consumption during the Outbreak,” Impact Covid-19 media Syst. Commun. Democr. consequences news Consum. Dur. outbreak, vol. 29, 2020, [Online]. Available: https://doi.org/10.3145/epi.2020.mar.23.
[12] D.Moskowitz, “Framing Responsibility for Political Issues: The Case of Public Assistance,” Polit. Behav., vol. 12, no. 1, pp. 19–40, 1990.
[13] W. R.Neuman, M. R.Just, andA. N.Crigler, “Common knowledge : news and the construction of political meaning,” 1992, [Online]. Available: https://doi.org/10.7208/chicago/9780226161174. 001.0001.
[14] S. N.Mutua andD. O.Ong’ong’a, “Online News Media Framing of COVID-19 Pandemic: Probing the Initial Phases of the Disease Outbreak in International Media,” Eur. J. Interact. Multimed. Educ., vol. 1, no. 2, 2020, doi: 10.30935/ejimed/8402.
[15] A. A.Olagoke, O. O.Olagoke, andA. M.Hughes, “Exposure to coronavirus news on mainstream media: The role of risk perceptions and depression,” Br. J. Health Psychol., vol. 25, no. 4, pp. 865–874, 2020, doi: 10.1111/bjhp.12427.
[16] M. K.Adu et al., “Prevalence and correlates of likely major depressive disorder among the adult population in ghana during the covid-19 pandemic,” Int. J. Environ. Res. Public Health, vol. 18, no. 13, pp. 1–19, 2021, doi: 10.3390/ijerph18137106.
[17] J.Torales et al., “COVID-19 infodemic and depressive symptoms: The impact of the exposure to news about COVID-19 on the general Paraguayan population,” J. Affect. Disord., vol. 298, no. October 2021, pp. 599–603, 2022, doi: 10.1016/j.jad.2021.11.036.
[18] B.Zarrouq et al., “An investigation of the association between religious coping, fatigue, anxiety and depressive symptoms during the COVID-19 pandemic in Morocco: a web-based cross-sectional survey,” BMC Psychiatry, vol. 21, no. 1, pp. 1–13, 2021, doi: 10.1186/s12888-021-03271-6.
[19] Y.Liu et al., “Aerodynamic analysis of SARS-CoV-2 in two Wuhan hospitals,” Nature, vol. 582, no. 7813, pp. 557–560, 2020, doi: 10.1038/s41586-020-2271-3.
[20] E.Jenelius andM.Cebecauer, “Impacts of COVID-19 on public transport ridership in Sweden: Analysis of ticket validations, sales and passenger counts,” Transp. Res. Interdiscip. Perspect., vol. 8, no. June, p. 100242, 2020, doi: 10.1016/j.trip.2020.100242.
[21] J.DeVos, “The effect of COVID-19 and subsequent social distancing on travel behavior,” Transp. Res. Interdiscip. Perspect., vol. 5, p. 100121, 2020, doi: 10.1016/j.trip.2020.100121.
[22] J. T.Wu, K.Leung, andG. M.Leung, “Nowcasting and forecasting the potential domestic and international spread of the 2019-nCoV outbreak originating in Wuhan, China: a modelling study,” Lancet, vol. 395, no. 10225, pp. 689–697, 2020, doi: 10.1016/S0140-6736(20)30260-9.
[23] C.Antoniou, G.Yannis, E.Papadimitriou, andS.Lassarre, “Relating traffic fatalities to GDP in Europe on the long term,” Accid. Anal. Prev., vol. 92, pp. 89–96, 2016, doi: 10.1016/j.aap.2016.03.025.
[24] E.Vingilis et al., “Coronavirus disease 2019: What could be the effects on Road safety?,” Accid. Anal. Prev., vol. 144, no. June, p. 105687, 2020, doi: 10.1016/j.aap.2020.105687.
[25] Y.Zhang andJ. D.Fricker, “Quantifying the impact of COVID-19 on non-motorized transportation: A Bayesian structural time series model,” Transp. Policy, vol. 103, no. January, pp. 11–20, 2021, doi: 10.1016/j.tranpol.2021.01.013.
[26] R. E.Park, “News as a Form of Knowledge: A Chapter in the Sociologyfile:///C:/Users/leo19/Desktop/A Content Analysis of Sexually Transmitted Diseases in the Print News Media.pdf of Knowledge,” Am. J. Sociol., vol. 45, no. 5, pp. 669–686, Mar.1940, doi: 10.1086/218445.
[27] D.McQuail, “Mass communication theory (5th edt.).” London: Sage Publications, 2005.
[28] J. R.Zaller, The nature and origins of mass opinion. Cambridge university press, 1992.
[29] C.Nwabueze andE.Okonkwo, “Rethinking the Bullet Theory in the Digital Age,” Int. J. Media, Journal. Mass Commun., vol. 4, no. 2, pp. 1–10, 2018, doi: 10.20431/2454-9479.0402001.
[30] P.Ene andJ.Ekpang, “The 2019 Nigerian Presidential Elections and the Resurgence of the Magic Bullet Theory of Media Effect,” Int. J. Commun. Soc., vol. 3, no. 2, pp. 52–62, 2021, doi: 10.31763/ijcs.v3i2.199.
[31] G. J.Rubin, H. W. W.Potts, andS.Michie, “The impact of communications about swine flu (influenza A HINIv) on public responses to the outbreak: Results from 36 national telephone surveys in the UK,” Health Technol. Assess. (Rockv)., vol. 14, no. 34, pp. 183–266, 2010, doi: 10.3310/hta14340-03.
[32] L.Zhang, Y.Kong, andH.Chang, “Media Use and Health Behavior in H1N1 Flu Crisis: The Mediating Role of Perceived Knowledge and Fear,” Atl. J. Commun., vol. 23, no. 2, pp. 67–80, 2015, doi: 10.1080/15456870.2015.1013101.
[33] P. L.Liu, “COVID-19 Information Seeking on Digital Media and Preventive Behaviors: The Mediation Role of Worry,” Cyberpsychology, Behav. Soc. Netw., vol. 23, no. 10, pp. 677–682, Jun.2020, doi: 10.1089/cyber.2020.0250.
[34] T. H.Chu, T. E. D.Yeo, andY.Su, “Effects of Exposure to COVID-19 News and Information: A Meta-Analysis of Media Use and Uncertainty-Related Responses During the Pandemic,” Journal. Mass Commun. Q., vol. 99, no. 1, pp. 89–112, 2022, doi: 10.1177/10776990211068857.
[35] B.Raghuram Kadali, S. S.Gadiraju, andH. K.Gaddam, “Analysis of Travelers’ Risk Perceptions About Public Transit Systems During COVID-19 Pandemic,” Transp. Res. Rec. J. Transp. Res. Board, p. 036119812210933, 2022, doi: 10.1177/03611981221093332.
[36] S.Shelat, T.Van DeWiel, E.Molin, J. W. C.vanLint, andO.Cats, “Analysing the impact of COVID-19 risk perceptions on route choice behaviour in train networks,” PLoS One, vol. 17, no. 3 March, pp. 1–19, 2022, doi: 10.1371/journal.pone.0264805.
[37] Z.Jin, Y.Yang, andY.Liu, “Stock closing price prediction based on sentiment analysis and LSTM,” Neural Comput. Appl., vol. 32, no. 13, pp. 9713–9729, 2020, doi: 10.1007/s00521-019-04504-2.
[38] F.Kollintza-Kyriakoulia, M.Maragoudakis, andA.Krithara, “Measuring the impact of financial news and social media on stock market modeling using time series mining techniques,” Algorithms, vol. 11, no. 11, 2018, doi: 10.3390/a11110181.
[39] J.Pirkis et al., “Suicide trends in the early months of the COVID-19 pandemic: an interrupted time-series analysis of preliminary data from 21 countries,” The Lancet Psychiatry, vol. 8, no. 7, pp. 579–588, 2021, doi: 10.1016/S2215-0366(21)00091-2.
[40] A. E.Davidson andL.Wallack, “A content analysis of sexually transmitted diseases in the print news media,” J. Health Commun., vol. 9, no. 2, pp. 111–117, 2004, doi: 10.1080/10810730490425268.
[41] C. B.Brown, “Tourism, crime and risk perception: An examination of broadcast media’s framing of negative Aruban sentiment in the Natalee Holloway case and its impact on tourism demand,” Tour. Manag. Perspect., vol. 16, pp. 266–277, 2015, doi: 10.1016/j.tmp.2014.12.001.
[42] J.Guo, J.Yang, Y.Zeng, andS.Zhu, “Please Let Science Talk: Presidential Sentiment and the Public’s Risk Perception in the Era of COVID-19.” Research Square, 2021, doi: 10.21203/rs.3.rs-954406/v1.
[43] Y.Zhang, J.Gao, X.Luo, X.Wu, andH.Bao, “Dynamic evolution of Public’s positive emotions and risk perception for the COVID-19 Pandemic: A case study of Hubei Province of China,” Math. Probl. Eng., vol. 2021, 2021, doi: 10.1155/2021/6680303.
[44] J.Dyer andB.Kolic, “Public risk perception and emotion on Twitter during the Covid-19 pandemic,” Appl. Netw. Sci., vol. 5, no. 1, 2020, doi: 10.1007/s41109-020-00334-7.
[45] P. J.Stone andE. B.Hunt, “A Computer Approach to Content Analysis: Studies Using the General Inquirer System,” in Proceedings of the May 21-23, 1963, Spring Joint Computer Conference, 1963, pp. 241–256, doi: 10.1145/1461551.1461583.
[46] P.Turney, “Thumbs Up or Thumbs Down? {S}emantic Orientation Applied to Unsupervised Classification of Reviews,” Comput. Res. Repos. - CORR, pp. 417–424, Dec.2002, doi: 10.3115/1073083.1073153.
[47] B.Pang, L.Lee, andS.Vaithyanathan, “Thumbs up? Sentiment Classification using Machine Learning Techniques,” in Proceedings of the 2002 Conference on Empirical Methods in Natural Language Processing ({EMNLP} 2002), Jul. 2002, pp. 79–86, doi: 10.3115/1118693.1118704.
[48] S.Wang, C.-R.Huang, Y.Yao, andA.Chan, “Word intuition agreement among Chinese speakers: a Mechanical Turk-based study,” Ling. Sin., vol. 3, no. 1, p. 13, 2017, doi: 10.1186/s40655-017-0032-5.
[49] A.Rogers, O.Kovaleva, andA.Rumshisky, “A primer in bertology: What we know about how bert works,” Trans. Assoc. Comput. Linguist., vol. 8, pp. 842–866, 2020, doi: 10.1162/tacl_a_00349.
[50] J.Devlin, M. W.Chang, K.Lee, andK.Toutanova, “BERT: Pre-training of deep bidirectional transformers for language understanding,” NAACL HLT 2019 - 2019 Conf. North Am. Chapter Assoc. Comput. Linguist. Hum. Lang. Technol. - Proc. Conf., vol. 1, no. Mlm, pp. 4171–4186, 2019.
[51] M. C.Buzea, S.Trausan-Matu, andT.Rebedea, “Automatic Fake News Detection for Romanian Online News,” Inf., vol. 13, no. 3, pp. 1–13, 2022, doi: 10.3390/info13030151.
[52] M. G.Sousa, K.Sakiyama, L. D. S.Rodrigues, P. H.Moraes, E. R.Fernandes, andE. T.Matsubara, “BERT for stock market sentiment analysis,” Proc. - Int. Conf. Tools with Artif. Intell. ICTAI, vol. 2019-Novem, pp. 1597–1601, 2019, doi: 10.1109/ICTAI.2019.00231.
[53] M.Shih, R.Duann, andS.Chung, “中文新聞文本之宣傳手法標記與分析,” Comput. Linguist. Chinese Lang. Process., vol. 26, no. 1, pp. 79–104, 2021.
[54] 謝吉隆、楊淳苾, “從「應變自然」到「社會應變」：以文字探勘方法檢視國內風災新聞的報導,” J. Educ. Media Libr. Sci., vol. 55, no. 3, pp. 1–24, Nov.2018, [Online]. Available: https://search.ebscohost.com/login.aspx?direct=true&db=lih&AN=133416511&lang=zh-tw&site=ehost-live.
[55] 林思平, “政治新聞與讀者政治立場的建構-以台灣三大本土報紙為例,” 復興崗學報, vol. 104, pp. 151–174, 2014.
[56] Z.Pan andG. M.Kosicki, “Framing analysis: An approach to news discourse,” Polit. Commun., vol. 10, no. 1, pp. 55–75, 1993, doi: 10.1080/10584609.1993.9962963.
[57] A. R. T.Schuck andA.Feinholdt, “News Framing Effects and Emotions,” Emerg. Trends Soc. Behav. Sci., no. May 2015, pp. 1–15, 2015, doi: 10.1002/9781118900772.etrds0241.
[58] G. E. P.Box andG. M.Jenkins, Time series analysis : forecasting and control, Rev. ed. San Francisco SE -: Holden-Day, 1976.
[59] R. B.Cleveland, W. S.Cleveland, J. E.McRae, andI.Terpenning, “STL: A seasonal-trend decomposition,” J. Off. Stat, vol. 6, no. 1, pp. 3–73, 1990.
[60] B. C.Monsell, “The X-13A-S Seasonal Adjustment Program Overview of X-13A-S New Modeling Options in X-13A-S,” 2007.
[61] C. C.Holt, “Forecasting seasonals and trends by exponentially weighted moving averages,” Int. J. Forecast., vol. 20, no. 1, pp. 5–10, 2004, doi: 10.1016/j.ijforecast.2003.09.015.
[62] U.S. Centers for Disease Control and Prevention (CDC), “Principles of epidemiology in public health practice : an introduction to applied epidemiology and biostatistics.” U.S. Department of Health and Human Services , Centers for Disease Control and Prevention, Atlanta, GA, 2012, doi: LK - https://worldcat.org/title/770946030.