運用BERT深度學習模型於衛教謠言檢測之研究

A Study on Using BERT to Identify Health Care Rumors

許文錦(Wen-Chin, Hsu)；李牧衡(Mu-Heng, Lee)；呂明聲(Ming-Sheng Lyu)

BERT ； 深度學習 ； 衛生教育 ； 謠言檢測 ； BERT ； deep learning technique ； health education ； rumor detection

資訊管理學報

29卷1期（2022 / 01 / 31）

27 - 44

繁體中文

近年來網路上流傳大量衛生教育謠言已為醫療照護人員帶來嚴重困擾，不實謠言除了誤導病患、造成恐慌外，更可能造成錯誤用藥、延誤治療等嚴重後果，更甚者，還可能造成醫病關係惡化，傷害醫師與醫院形象, 因此發展新的闢謠與防範機制成為急需研究的重要議題。關於謠言防範，醫療機構雖提供正確衛教宣導，但侷於人力與資源不足，成效有限，而人工智慧技術快速發展為此問題帶來新的解決方向。本研究運用深度學習技術BERT（Bidirectional Encoder Representations from Transformers）發展新一代衛教謠言檢測系統。實作上，首先收集包括衛福部等15個政府與民間單位衛教謠言資料，經資料前處理後，運用BERT結合Bidirectional Long Short-Term Memory（BiLSTM）進行模型訓練與驗證，最後將此BERT-BiLSTM闢謠模型佈署於Line聊天機器人上供使用者試用。實驗結果顯示本研究提出之BERT-BiLSTM模式可準確辨識出90%的衛教謠言，本研究成果將有助於緩解醫護人員衛教解說時間不足、與滿足民眾查證衛教資訊之需求。

The circulation of numerous health care rumors on the internet in recent years has become extremely troublesome for medical and healthcare personnel. False rumors have misled patients, created panic, and even resulted in medication misuse and delayed treatment. They may also cause deteriorating doctor-patient relationships and tarnish the images of doctors and hospitals. Developing new mechanisms to dispel and prevent rumors has thus become an imperative issue in research. Although medical institutions promote accurate health education, limited labor and resources limit their effectiveness. The rapid progress in artificial intelligence technology is offering a new direction to solve this problem. This study employed BERT (Bidirectional Encoder Representations from Transformers), a deep learning technique, to develop the new generation alert system for health care rumors. First of all, we collected the materials of health care rumors from 15 private organizations and government agencies, including the Ministry of Health and Welfare. After data preprocessing was done, we used BERT combined with Bidirectional Long Short-Term Memory (BiLSTM) to conduct a model training and validation. Lastly, we deployed the BERT-BiLSTM model in Line chatbot so some LINE users can try it. The experiment result showed that the proposed model can identify as high as 90% of the health care rumors in social media. The research results will help to alleviate the burden relating to healthcare promotion by health professionals and to meet the public’s needs to search for the correct health information.

1. Lo, W.L.,Chiu, M.H.P.(2015).A Content Analysis of Internet Health Rumors.Journal of Educational Media & Library Sciences,52(1),1-24.
   連結：
2. Ajao, O.,Bhowmik, D.,Zargari, S.(2018).Fake news identification on twitter with hybrid cnn and rnn models.Proceedings of the 9th International Conference on Social Media and Society
3. Asghar, M. Z.,Habib, A.,Habib, A.,Khan, A.,Ali, R.,Khattak, A.(2019).Exploring deep neural networks for rumor detection.Journal of Ambient Intelligence and Humanized Computing,12,4315-4333.
4. Bastable, S. B.(2016).Essentials of patient education.MA:Jones & Bartlett Publishers.
5. Bondielli, A.,Marcelloni, F.(2019).A survey on fake news and rumour detection techniques.Information Sciences,497,38-55.
6. Bordia, P.,DiFonzo, N.(2017).Psychological motivations in rumor spread.Rumor Mills: The Social Impact of Rumor and Legend
7. Cao, J.,Guo, J.,Li, X.,Jin, Z.,Guo, H.,Li, J.(2018).,未出版
8. Castillo, C.,Mendoza, M.,Poblete, B.(2011).Information credibility on twitter.Proceedings of the 20th international conference on World wide web
9. Devlin, J.,Chang, M.W.,Lee, K.,Toutanova, K.(2018).,未出版
10. DiFonzo, N.,Bordia, P.(2000).How top PR professionals handle hearsay: Corporate rumors, their effects, and strategies to manage them.Public Relations Review,26(2),173-190.
11. DiFonzo, N.,Robinson, N. M.,Suls, J. M.,Rini, C.(2012).Rumors about cancer: Content, sources, coping, transmission, and belief.Journal of health communication,17(9),1099-1115.
12. Goh, D. H.L.,Chua, A. Y.,Shi, H.,Wei, W.,Wang, H.,Lim, E. P.(2017).An analysis of rumor and counter-rumor messages in social media.Proceedings of 19th International Conference on Asia-Pacifc Digital Libraries
13. Huilai, Z.,Fang, L.,Junjie, Z.(2016).What makes people resend healthy food messages online: the effects of message cues.Proceedings of 2016 13th International Conference on Service Systems and Service Management (ICSSSM)
14. Jawahar, G.,Sagot, B.,Seddah, D.(2019).What does BERT learn about the structure of language?.ACL 2019-57th Annual Meeting of the Association for Computational Linguistics
15. Jordan, M. I.(1995).MIT Computational Cognitive Science ReportMIT Computational Cognitive Science Report,未出版
16. Kwon, S.,Cha, M.,Jung, K.,Chen, W.,Wang, Y.(2013).Prominent features of rumor propagation in online social media.Proceeding of 2013 IEEE 13th International Conference on Data Mining
17. Li, B.,Chong, A.(2019).What Influences the Dissemination of Online Rumor Messages: Message Features and Topic-congruence.Proceedings of the 40th International Conference on information systems
18. Li, J.(2019).Detecting False Information in Medical and Healthcare Domains: A Text Mining Approach.Proceeding of International Conference on Smart Health
19. Li, L.,Cai, G.,Chen, N.(2018).A Rumor Events Detection Method Based on Deep Bidirectional GRU Neural Network.Proceeding of 2018 IEEE 3rd International Conference on Image, Vision and Computing (ICIVC)
20. Liang, J.,Yang, M.(2015).On spreading and controlling of online rumors in wemedia era.Asian Culture and History,7(2),42-46.
21. Ma, J.,Gao, W.,Mitra, P.,Kwon, S.,Jansen, B. J.,Wong, K.F.,Cha, M.(2016).Detecting rumors from microblogs with recurrent neural networks.Proceedings of the 25th International Joint Conference on Artificial Intelligence (IJCAI 2016)
22. Ma, J.,Gao, W.,Wong, K.F.(2019).Detect Rumors on Twitter by Promoting Information Campaigns with Generative Adversarial Learning.Proceedings of The World Wide Web Conference 2019
23. Matthews, A. K.,Sellergren, S. A.,Manfredi, C.,Williams, M.(2002).Factors influencing medical information seeking among African American cancer patients.Journal of health communication,7(3),205-219.
24. Moein, S.(2014).Medical diagnosis using artificial neural networks.IGI Global.
25. Nguyen, T. N.,Li, C.,Niederée, C.(2017).On early-stage debunking rumors on twitter: Leveraging the wisdom of weak learners.Proceeding of International Conference on Social Informatics
26. Nutbeam, D.(2008).The evolving concept of health literacy.Social Science & Medicine,67(12),2072-2078.
27. Nwankpa, C.,Ijomah, W.,Gachagan, A.,Marshall, S.(2018).,未出版
28. Oh, H. J.,Lee, H.(2019).When Do People Verify and Share Health Rumors on Social Media? The Effects of Message Importance, Health Anxiety, and Health Literacy.Journal of health communication,24(11),837-847.
29. Paek, H.-J.,Hove, T.(2019).Effective strategies for responding to rumors about risks: The case of radiation-contaminated food in South Korea.Public Relations Review,45(3),1-9.
30. Park, Y.-J.,Bae, J. H.,Shin, M. H.,Hyun, S. H.,Cho, Y. S.,Choe, Y. S.,Choi, J. Y.,Lee, K.-H.,Kim, B.-T.,Moon, S. H.(2019).Development of Predictive Models in Patients with Epiphora Using Lacrimal Scintigraphy and Machine Learning.Nuclear medicine and molecular imaging,53(2),125-135.
31. Pendleton, S. C.(1998).Rumor research revisited and expanded.Language & Communication,18(1),69-86.
32. Sankar, H.,Subramaniyaswamy, V.,Vijayakumar, V.,Arun Kumar, S.,Logesh, R.,Umamakeswari, A.(2019).Intelligent sentiment analysis approach using edge computing‐based deep learning technique.Software: Practice and Experience,50(5),645-657.
33. Sharma, M.(2016).Theoretical foundations of health education and health promotion.Jones & Bartlett Publishers.
34. Singh, J. P.,Rana, N. P.,Dwivedi, Y. K.(2019).Rumour Veracity Estimation with Deep Learning for Twitter.Proceeding of International Working Conference on Transfer and Diffusion of IT
35. Sjöström, A. E.,Hörnsten, Å.,Hajdarevic, S.,Emmoth, A.,Isaksson, U.(2019).Primary Health Care Nurses’ Experiences of Consultations With Internet-Informed Patients: Qualitative Study.JMIR Nursing,2(1),e14194.
36. Vaswani, A.,Shazeer, N.,Parmar, N.,Uszkoreit, J.,Jones, L.,Gomez, A. N.,Kaiser, Ł.,Polosukhin, I.(2017).Attention is all you need.Proceeding of Advances in neural information processing systems
37. Vosoughi, S.,Roy, D.,Aral, S.(2018).The spread of true and false news online.Science,359(6380),1146-1151.
38. Wang, Y.,McKee, M.,Torbica, A.,Stuckler, D.(2019).Systematic literature review on the spread of health-related misinformation on social media.Social Science & Medicine,240,112552.
39. Wolf, T.,Debut, L.,Sanh, V.,Chaumond, J.,Delangue, C.,Moi, A.,Cistac, P.,Rault, T.,Louf, R.,Funtowicz, M.(2019).,未出版
40. Yang, F.,Liu, Y.,Yu, X.,Yang, M.(2012).Automatic detection of rumor on Sina Weibo.Proceedings of the ACM SIGKDD Workshop on Mining Data Semantics
41. Yu, F.,Liu, Q.,Wu, S.,Wang, L.,Tan, T.(2017).A convolutional approach for misinformation identification.Proceeding of IJCAI
42. Zaccone, G.,Karim, M. R.(2018).Deep Learning with TensorFlow: Explore neural networks and build intelligent systems with Python.Packt Publishing Ltd.
43. Zhao, Z.,Resnick, P.,Mei, Q.(2015).Enquiring minds: Early detection of rumors in social media from enquiry posts.Proceedings of the 24th International Conference on World Wide Web
44. Zhou, Z.,Qi, Y.,Liu, Z.,Yu, C.,Wei, Z.(2018).A C-GRU Neural Network for Rumors Detection.Proceeding of 2018 5th IEEE International Conference on Cloud Computing and Intelligence Systems (CCIS)
45. Zubiaga, A.,Liakata, M.,Procter, R.,Wong Sak Hoi, G.,Tolmie, P.(2016).Analysing how people orient to and spread rumours in social media by looking at conversational threads.PloS one,11(3),e0150989.
46. 汪志堅,駱少康(2002)。以內容分析法探討網路謠言之研究。資訊、科技與社會學報，1，131-149。
47. 陳文信 (2010) 。「醫師看診平均 1.4 分鐘監院糾正」。 中國時報，http://healthmedia.nownews.com/contents.aspx?cid=5,69&id=10404。
48. 曾芬郁,郭瑞祥(2009)。「門診時段及醫師特質對就診流程品質之影響。台灣醫學，16(6)，558-566。

1. (2024)。聊天機器人不確定性與信任、品牌參與之關係。東吳經濟商學學報，108，37-73。