在產學研界的努力下，人工智慧崛起與大數據平台日漸發展，帶動產業型態的改變。人工智慧在醫療上的應用主要是提供客觀的診斷標準和減少醫療支出，發展高效率、精確且穩定的協助工具。本論文研究深度學習在疾病預測之應用。
本論文第一部分研究基於影像前處理之深度學習(Deep Learning)預測糖尿病視網膜病變，針對眼底影像使用卷積類神經網路 (Convolutional Neural Network, CNN)進行自動偵測視網膜病變，使用原始眼底影像前處理之影像熵(Entropy Images)，增強視網膜病變與周邊結構的對比性，提高影像特徵的顯著性，因此增加預測的正確性。我們從公開的網路資料上獲取可判讀的原始視網膜照片，所有的原始視網膜照片前處理成影像熵，原始影像和影像熵為卷積類神經網路輸入影像，並比較兩組影像分別偵測視網膜病變存在和視網膜中重度嚴重病變存在的正確性。實驗結果顯示影像熵相對於原始影像深度學習後，分別偵測視網膜病變及視網膜中重度嚴重病變存在的精確度、敏感度和特異度，統計上有明顯的提升，此研究成果可提高糖尿病視網膜病變診斷的精確度。
本論文第二部分研究以空氣品質監測數據來預測上呼吸道感染的門診人數之趨勢，使用網路公開的大數據資料，針對PM2.5和PM10空氣懸浮微粒濃度，使用深度學習來預測上呼吸道感染病患的來院門診人數之趨勢。我們使用行政院環境保護署PM2.5和PM10的濃度資料和衛生福利部疾病管制署上呼吸道門診就診人數，利用多層感知器模型(Multilayer Perceptron, MLP)，使用連續30天PM濃度做深度學習，預測下一週門診上呼吸道感染量(三等分為高、中、低量)的精確度。實驗結果顯示對於台灣總人口，PM2.5和PM10預測精確度分別是81.75%和83.21%，對65歲以上老年族群為89.05% 和 88.32%，本論文並分析台灣各區域之預測精確度。此研究成果可提供上呼吸道門診人數之趨勢預測。
本論文研究主題在醫療應用，可開拓人工智慧在醫療業的新視野，期許能提升國家科技發展的競爭力。

The academics and industrials have made great efforts on artificial intelligence based on a variety of big data platforms establishment. Reciprocally, the emergence of artificial intelligence has changed the industrials. Using artificial intelligence could provide an objective diagnosis of disease, reduce the cost, and increase high efficient, accurate, and stable tool for medical practice. This dissertation investigates the applications of deep learning on disease prediction.
In the first part of this dissertation, the Convolutional Neural Network (CNN) is used for the deep learning of diabetic retinopathy (DR) images. The method is focused on the pre-processing of fundus photographs using entropy images, which represent the complexity of original fundus photographs. Entropy images may strengthen the contrast between DR lesions and unaffected areas and thus increase the detection accuracy. A large sample of interpretable fundus photographs is obtained from a publicly available data set and all photographs are transformed into entropy images. Both the original fundus photographs and the entropy images are used as the inputs of CNN, and the results of detecting any DR and referable DR as the outputs from the two data sets are compared. The results show that the detection accuracy, sensitivity and specificity for any DR or referable DR of using the entropy images with CNN training are statistical significantly better than those of the original photographs. The research results can increase the diagnostic accuracy to detect the severity of DR automatically.

摘要.............................................i
Abstract.......................................iii
Table of Contents...............................vi
List of Figures...............................viii
List of Tables...................................x
Chapter 1 Introduction...........................1
1.1 Deep Learning and the Applications on Disease Prediction .................................................3
1.2 Motivation..............................5
1.3 Dissertation Organization...............7
Chapter 2 Entropy Images in Deep Learning for Diabetic Retinopathy .................................................8
2.1 Pre-processing of Fundus Photography....8
2.2 Materials and Methods..................11
2.2.1 Data Sets...............................11
2.2.2 Grading.................................11
2.2.3 Pre-processing Images...................14
2.2.4 Entropy Images..........................16
2.2.5 Deep Feature Learning...................18
2.3 Experimental Results...................19
2.3.1 Statistical Analysis......................19
2.3.2 Performance Comparison....................20
2.4 Discussion..............................25
Chapter 3 Deep Learning of PM levels on Prediction of Upper Respiratory Tract Infections....................29
3.1 Background.............................29
3.2 Data Collection........................30
3.3 Method and Experiments.................35
3.3.1 Volume of Outpatient Visits for URI.....35
3.3.2 Multilayer Perceptron (MLP) Model .......39
3.4 Experimental Results........................41
3.4.1 Baseline PM and Outpatient Data...........41
3.4.2 Results...................................53
3.5 Discussion..............................58
Chapter 4 Conclusions and Future Work...........63
4.1 Conclusion of the First Part............63
4.2 Conclusion of the Second Part...........64
4.3 Future Work.............................64
References......................................66

[1]F. Jiang, Y. Jiang, H. Zhi, Y. Dong, H. Li, S. Ma, Y. Wang, Q. Dong, H. Shen, and Y. Wang, “Artificial intelligence in healthcare: past, present and future,” Stroke and Vascular Neurology, vol. 2, no. 4, pp. 230-243, June 2017.
[2]K. Yasaka, H. Akai, O. Abe, and S. Kiryu, “Deep learning with convolutional neural network for differentiation of liver masses at dynamic contrast-enhanced CT: A preliminary study,” Radiology, vol. 286, no. 3, pp. 887-896. March 2018.
[3]G. Kang, K. Liu, B. Hou, and N. Zhang, “3D multi-view convolutional neural networks for lung nodule classification,” PLoS One, vol. 12, no. 11, e0188290, November 2017.
[4]X. Zhang, L.F. Yan, Y.C. Hu, G. Li, Y. Yang, Y. Han, Y.Z. Sun, Z.C. Liu, Q. Tian, Z.Y. Han, L.D. Liu, B.Q. Hu, Z.Y. Qiu, W. Wang, and G.B. Cui, “Optimizing a machine learning based glioma grading system using multi-parametric MRI histogram and texture features,” Oncotarget, vol. 8, no. 29, pp.47816-47830, July 2017.
[5]L.M. Prevedello, B.S. Erdal, J.L. Ryu, K.J. Little, M. Demirer, S. Qian, and R.D. White, “Automated critical test findings identification and online notification system using artificial intelligence in imaging,” Radiology, vol. 285, no. 3, pp. 923-931, December 2017.
[6]C. Cortes and V. Vapnik, “Support-vector networks,” Machine Learning, vol. 20, pp. 273–293, 1995.
[7]J.R. Quinlan, “Induction of decision trees,” Machine Learning, vol. 1, pp. 81-106, 1986.
[8]T.K. Ho, “The random subspace method for constructing decision forests,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 20, no. 8, pp. 832-844, August 1998.
[9]P. Domingos and M. Pazzani, “On the optimality of the simple Bayesian classifier under zero-one loss,” Machine Learning, vol. 29, pp. 103–137, 1997.
[10]S.B. Choi, W. Lee, J.H. Yoon, J.U. Won, and D.W. Kim, “Ten-year prediction of suicide death using Cox regression and machine learning in a nationwide retrospective cohort study in South Korea,” Journal of Affective Disorders, vol. 231, pp. 8-14, April 2018.
[11]B. Ambale-Venkatesh, X. Yang, C.O. Wu, K. Liu, W.G. Hundley, R. McClelland, A.S. Gomes, A.R. Folsom, S. Shea, E. Guallar, D.A. Bluemke, and J.A.C. Lima, “Cardiovascular event prediction by machine learning: the multi-ethnic study of atherosclerosis,” Circulation Research, vol. 121, no. 9, pp. 1092-1101, October 2017.
[12]R. Shouval, A. Hadanny, N. Shlomo, Z. Iakobishvili, R. Unger, D. Zahger, R. Alcalai, S. Atar, S. Gottlieb, S. Matetzky, I. Goldenberg, and R. Beigel, “Machine learning for prediction of 30-day mortality after ST elevation myocardial infraction: an acute coronary syndrome Israeli survey data mining study,” International Journal of Cardiology, vol. 246, pp. 7-13, November 2017.
[13]C. Chrysostomou, H. Partaourides, and H. Seker, “Prediction of influenza A virus infections in humans using an artificial neural network learning approach,” in Proceedings of the 39th Annual International Conference of IEEE Engineering in Medicine and Biology Society, pp. 1186-1189, July 2017.
[14]V. Gulshan, L. Peng, M. Coram, M.C. Stumpe, D. Wu, A. Narayanaswamy, S. Venugopalan, K. Widner, T. Madams, J. Cuadros, R. Kim, R. Raman, P.C. Nelson, J.L. Mega, and D.R. Webster, “Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs,” Journal of the American Medical Association, vol. 316, no. 22, pp. 2402-2410, December 2016.
[15]D.S.W. Ting, C.Y. Cheung, G. Lim, G.S.W. Tan, N.D. Quang, A. Gan, H. Hamzah, R. Garcia-Franco, I.Y. San Yeo, S.Y. Lee, E.Y.M. Wong, C. Sabanayagam, M. Baskaran, F. Ibrahim, N.C. Tan, E.A. Finkelstein, E.L. Lamoureux, I.Y. Wong, N.M. Bressler, S. Sivaprasad, R. Varma, J.B. Jonas, M.G. He, C.Y. Cheng, G.C.M. Cheung, T. Aung, W. Hsu, M.L. Lee, and T.Y. Wong. “Development and validation of a deep learning system for diabetic retinopathy and related eye diseases using retinal images from multiethnic populations with diabetes,” Journal of the American Medical Association, vol. 318, no. 22, pp. 2211-2223, December 2017.
[16]R. Gargeya and T. Leng. “Automated identification of diabetic retinopathy retinopathy using deep learning,” Ophthalmology, vol. 124, no. 7, pp. 962-969, July 2017.
[17]G.M. Lin, M.J. Chen, Y.Y. Lin, M.H. Lin, and C.H. Yeh, “An improvement of machinery detection for any diabetic retinopathy by pre-processing retinal photographs to entropy images in deep learning,” in Proceedings of the 33rd Asia-Pacific Academy of Ophthalmology (APAO) Congress, Hong Kong, February 2018.
[18]“https://www.electronicproducts.com/Programming/Software/The_CDC_uses_machine_learning_and_social_media_to_forecast_flu_outbreaks.aspx,” accessed on March 19, 2018.
[19]J.P. Parreco, A.E. Hidalgo, A.D. Badilla, O. Ilyas and R. Rattan, “Predicting central line-associated bloodstream infections and mortality using supervised machine learning,” Journal of Critical Care, vol. 45, pp. 156-162, February 2018.
[20]X. Zhang, J.B Saaddine, C.F. Chou, M.F. Cotch, Y.J. Cheng, L.S. Geiss, E.W. Gregg, A.L. Albright, B.E. Klein, and R. Klein, “Prevalence of diabetic retinopathy in the United States, 2005-2008,” Journal of the American Medical Association, vol. 304, no. 6, pp. 649-656, August 2010.
[21]P.D. Jani, L. Forbes, A. Choudhury, J.S. Preisser, A.J. Viera, and S. Garg. Evaluation of diabetic retinal screening and factors for ophthalmology referral in a telemedicine network,” JAMA Ophthalmology, vol. 135, no. 7, pp.706-714, July 2017.
[22]L. Sellahewa, C. Simpson, P. Maharajan, J. Duffy, and I. Idris, “Grader agreement, and sensitivity and Specificity of digital photography in a community optometry-based diabetic eye screening program,” Clinical Ophthalmology, vol. 8, pp.1345-1349, July 2014.
[23]P. Ruamviboonsuk, N. Wongcumchang, P. Surawongsin, E. Panyawatananukul, and M. Tiensuwan, “Screening for diabetic retinopathy in rural area using single-field, digital fundus images,” Journal of the Medical Association of Thailand, vol. 88, no. 2, pp. 176-180, 2015.
[24]L.Z. Wang, C.Y. Cheung, R.J. Tapp, H. Hamzah, G. Tan, D. Ting, E. Lamoureux, and T.Y. Wong, “Availability and variability in guidelines on diabetic retinopathy screening in Asian countries,” British Journal of Ophthalmology, vol. 101, no. 10, pp. 1352-1360, October 2017.
[25]T.J. Grahame, R. Klemm, and R.B. Schlesinger, “Public health and components of particulate matter: the changing assessment of black carbon,” Journal of the Air & Waste Management Association, vol. 64, no. 6, pp. 620-660, November 2014.
[26]M.A. Bind, A. Baccarelli, A. Zanobetti, L. Tarantini, H. Suh, P. Vokonas, and J. Schwartz, “Air pollution and markers of coagulation, inflammation, and endothelial function: associations and epigene-environment interactions in an elderly cohort,” Epidemiology, vol. 23, no. 2, pp. 332-340, March 2014.
[27]Y. Liao, L. Xu, X. Lin, and Y.T. Hao, “Temporal trend in lung cancer burden attributed to ambient fine particulate matter in Guangzhou, China,” Biomedical and Environmental Sciences, vol. 30, no. 10, pp. 708-717, October 2017.
[28]Y. Zhou, L. Li, and L. Hu. “Correlation analysis of PM10 and the incidence of lung cancer in Nanchang, China,” International Journal of Environmental Research and Public Health, vol. 14, no. 10, p. E1253, October 2017.
[29]R. Li, N. Jiang, Q. Liu, J. Huang, X. Guo, F. Liu, and Z. Gao, “Impact of air pollutants on outpatient visits for acute respiratory outcomes,” International Journal of Environmental Research and Public Health, vol.14, no. 1, p. E47, January 2017.
[30]A.H. Sinclair, E.S. Edgerton, R. Wyzga, and D. Tolsma, “A two-time-period comparison of the effects of ambient air pollution on outpatient visits for acute respiratory illnesses,” Journal of the Air & Waste Management Association, vol. 60, no. 2, pp. 163-175, 2010.
[31]A. Faustini, M. Stafoggia, P. Colais, G. Berti, L. Bisanti, E. Cadum, A. Cernigliaro, S. Mallone, C. Scarnato, F. Forastiere, and EpiAir Collaborative Group, “Air pollution and multiple acute respiratory outcomes,” European Respiratory Journal, vol. 42, no. 2, pp. 304-313, August 2013.
[32]E. Decencière, G. Cazuguel, X. Zhang, G. Thibault, J.C. Klein, F. Meyer, B. Marcotegui, G. Quellec, M. Lamard, R. Danno, D. Elie, P. Massin, Z. Viktor, A. Erginay, B. Lay, and A. Chabouis, “TeleOphta: machine learning and image processing methods for teleophthalmology,” Innovation and Research in BioMedical Engineering, vol. 34, no. 2, pp.196-203, April 2013.
[33]G. Quellec, M. Lamard, P.M. Josselin, G. Cazuguel, B. Cochener, and C. Roux, “Optimal wavelet transform for the detection of microaneurysms in retina photographs,” IEEE Transactions on Medical Imaging, vol. 27, no. 9, pp. 1230-1241, September 2008.
[34]R. Gargeya and T. Leng, “Automated identification of diabetic retinopathy using deep learning,” Ophthalmology, vol. 124, no. 7, pp.962-969, July 2017.
[35]G. Quellec, K. Charrière, Y. Boudi, B. Cochener, and M. Lamard, “Deep image mining for diabetic retinopathy screening,” Medical Image Analysis, vol. 39, pp.178-193, July 2017.
[36]W.J. Huang, C.H. Yeh, C.C. Kuo, Y.C. Cheng, and J.Y. Lin, “Rating realism assessment for computer generated imagery,” in Proceedings of 2017 IEEE International Conference on Consumer Electronics - Taiwan (ICCE-TW), June 2017.
[37]“https://www.kaggle.com/c/diabetic-retinopathy-detection,” accessed on November 1, 2017.
[38]M.E. T Chandrakumar and M.E. R Kathirvel, “Classifying diabetic retinopathy using deep learning architecture,” International Journal of Engineering Research & Technology, vol. 5, no. 6, pp. 19-24, June 2016.
[39]W.M. Gondal, J.M. Köhler, R. Grzeszick, G.A. Fink, and M. Hirsch, “Weakly-supervised localization of diabetic retinopathy lesions in retinal fundus images,” in Proceedings of IEEE International Conference on Image Processing (ICIP), Bejing, China, September 2017.
[40]M.D. Abràmoff, Y. Lou, A. Erginay, W. Clarida, R. Amelon, J.C. Folk, and M. Niemeijer, “Improved automated detection of diabetic retinopathy on a publicly available dataset through integration of deep learning,” Investigative Ophthalmology & Visual Science, vol. 57, no. 13, pp. 5200-5206, October 2016.
[41]American Academy of Ophthalmology. International clinical diabetic retinopathy disease severity scale detailed table. “http://www.icoph.org/dynamic/attachments/resources/diabetic-retinopathy-detail.pdf,” accessed on November 1, 2017.
[42]L. Liu, B. Liu, H. Huang, and A.C. Bovik,“No-Reference image quality assessment based on spatial and spectral entropies,” Signal Processing: Image Communication, vol. 29, no. 8, pp. 856-863, September 2014.
[43]F. Yu, J. Sun, A. Li, J. Cheng, C. Wan, and J. Liu, “Image quality classification for DR screening using deep learning,” in Proceedings of the 39th Annual International Conference of IEEE Engineering in Medicine and Biology Society, pp. 664-667, July 2017.
[44]Y. Morales, R. Nuñez, J. Suarez, and C. Torres, “Digital tool for detecting diabetic retinopathy in retinography image using Gabor transform,” Journal of Physics: Conference Series, vol. 792, no. 1, p. 012083, 2017.
[45]B. Antal and A. Hajdu, “An ensemble-based system for automatic screening of diabetic retinopathy,” Knowledge-Based Systems, vol. 60, pp.20-27, 2014.
[46]C.J. Sánchez, M. Niemeijer, A.V. Dumitrescu, M.S. Suttorp-Schulten, M.D. Abràmoff, and B. van Ginneken. “Evaluation of a computer-aided diagnosis system for diabetic retinopathy screening on public data,” Investigative Ophthalmology & Visual Science, vol. 52, no. 7, pp. 4866-4871, July 2011.
[47]L. Seoud, T. Hurtut, J. Chelbi, F. Cheriet, and J.M. Pierre Langlois, “Red lesion detection using dynamic shape features for diabetic retinopathy screening,” IEEE Transactions on Medical Imaging, vol. 35, no. 4, pp. 1116-1126, April 2016.
[48]S. Roychowdhury, D.D. Koozekanani, and K.K. Parhi, “DREAM: diabetic retinopathy analysis using machine learning,” IEEE Journal of Biomedical and Health Informatics, vol. 18, no. 5, pp. 1717-1728, September 2014.
[49]I. Djalalova, L.D. Monache, and J. Wilczak, “PM 2.5 analog forecast and kalman filter post-processing for the community multiscale air quality (cmaq) model,” Atmospheric Environment, vol. 119, pp. 431–442, May 2015.
[50]B. Lv, Y. Liu, P. Yu, B. Zhang, and Y. Bai, “Characterizations of PM2.5 pollution pathways and sources analysis in four large cities in China,” Aerosol and Air Quality Research, vol. 15, pp. 1836–1843, 2015.
[51]W. Sun, H. Zhang, A. Palazoglu, A. Singh, W. Zhang, and S. Liu, “Prediction of 24-hour-average PM 2.5 concentrations using a hidden Markov model with different emission distributions in Northern California,” Science of the Total Environment, vol. 443, pp. 93–103, January 2013.
[52]G. Shangzan, X. Da, and Y. Xingyuan, “A short-term rainfall prediction method based on neural networks and model ensemble,” Advances in Meteorological Science and Technology, vol.7, no. 1, pp. 107–113, 2017.
[53]X. Shi, Z. Chen, H. Wang, D.Y. Yeung, W.K. Wong, and W.C. WOO, “Convolutional LSTM network: a machine learning approach for precipitation nowcasting,” in Proceedings of the 28th International Conference on Neural Information Processing Systems, vol 1, pp. 802-810, December 2015.
[54]B. T. Ong, K. Sugiura, and K. Zettsu, “Dynamically pre-trained deep recurrent neural networks using environmental monitoring data for predicting PM2.5,” Neural Computing and Applications, vol. 27, no. 6, pp. 1553–1566, August 2016.
[55]X. Liang, S. Li, S. Zhang, H. Huang, and S.X. Chen, “PM2.5 data reliability, consistency, and air quality assessment in five Chinese cities,” Journal of Geophysical Research: Atmospheres, vol. 121, no. 17, pp. 10220–10236, September 2016.
[56]J. Tang, C. Deng, and G.B. Huang, “Extreme learning machine for multilayer perceptron,” IEEE Transactions on Neural Networks and Learning Systems, vol. 27, no. 4, pp. 809-821, April 2016.
[57]“https://taqm.epa.gov.tw/taqm/en/,” accessed on November 1, 2017.
[58]“https://data.cdc.gov.tw/,” accessed on November 1, 2017.
[59]J. Wang, S and Ogawa, “Effects of meteorological conditions on PM2.5 concentrations in Nagasaki, Japan,” International Journal of Environmental Research and Public Health, vol. 12, no. 8, pp. 9089-9101, August 2015.
[60]J. Feng, H. Yu, K. Mi, X. Su, Y. Li, Q. Li, and J. Sun, “One year study of PM2.5 in Xinxiang city, North China: seasonal characteristics, climate impact and source,” Ecotoxicology and Environmental Safety, vol. 154, pp.75-83, June 2018.
[61]J.J. Jaakkola, M. Paunio, M. Virtanen, and O.P. Heinonen, “Low-level air pollution and upper respiratory infections in children,” American Journal of Public Health, vol. 81, no. 8, pp.1060-1063, August 1991.
[62]N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhutdinov, “Dropout: a simple way to prevent neural networks from overfitting,” Journal of Machine Learning Research, vol. 15, pp. 1929-1958, June 2014.