臺灣博碩士論文加值系統

隨著科技的進步以及醫療水準的提升，台灣整體社會人口已經呈現高齡化的趨勢。高齡化人口的死亡原因當中又以慢性病為主要因素，各項慢性病中又以慢性阻塞性肺病對高齡患者尤為嚴重，而胃食道逆流為常見的大眾疾病。慢性阻塞性肺病以抽菸族群最為常見，台灣抽菸人口眾多，嚴重影響國人的健康，但對這項慢性病卻少有聽聞以及相關的研究探討過其與胃食道逆流之間的關係。
故本研究以國內醫療機構相關資料庫為研究，篩選出罹患慢性阻塞性肺病之病患，並運用基因邏輯斯迴歸演算法、粒子群最佳化演算法以及交叉熵演算法計算各項因子的權重並與倒傳遞類神經網路及支援向量機相互結合，建立六種預測胃食道逆流之風險模型，並結合案例式推理系統，評估病患是否會有伴隨胃食道逆流的風險，並設計評估介面，方便醫療人員於臨床上進行併發症的風險評估。
各預測模型經K疊交互驗證結合網格搜索法進行參數調校後，模型效能皆有 85%以上的分類準確度，ROC曲線下面積皆為0.879以上。經傅立曼檢定發現，預測模型之間在分類準確度與ROC曲線下面積，兩效能衡量指標下，皆存在顯著效能差異，因此各模型進一步使用成對樣本T檢定預測模型之個別優劣性。在分類準確度衡量基礎下，基因邏輯斯迴歸結合倒傳遞類神經網路與粒子群演算法結合支援向量機和粒子群演算法結合倒傳遞類神經網路之模型效能表現同等優異，最高準確度為95.16%；在ROC曲線下面積衡量基礎下，基因邏輯斯迴歸結合倒傳遞類神經網路與粒子群演算法結合倒傳遞類神經網路和基因邏輯斯迴歸結合支援向量機之模型效能表現同等優異，最高ROC曲線下面積達0.95803092。評估系統經K疊交互驗證評估其效能，各模型皆有92.85%以上分類準確度，ROC曲線下面積皆有0.928以上。經傅立曼檢定發現，評估系統在導入上述三類演算法權重值下，在分類準確度與ROC曲線下面積，兩效能衡量指標下，皆不存在顯著效能差異；因此，皆適合作為評估系統之權重運算，本研究可提供相關醫療機構做為預測評估之參考依據。

With the advancement of science and technology and the improvement of medical standards, Taiwan's overall social population has shown a trend of aging. Chronic diseases are the main cause of death among the aging population, Chronic Obstructive Pulmonary Disease(COPD) is particularly serious in elderly patients, and Gastroesophageal Reflux Diseases(GERD) are very common in life. COPD is most common among smokers. There is a large smoking population in Taiwan, which seriously affects the health of Chinese people. However, little has been heard about this chronic disease and related studies have explored the relationship between COPD and GERD.
This research in domestic medical institutions related to the database for research selected patients suffering from a chronic obstructive pulmonary disease (COPD), and use a Genetic Logistic Regression Algorithm(GALR), Particle Swarm Optimization Algorithm(PSO), and Cross-Entropy Algorithm(CE) to calculate the weight of each factor and transfer and Back Propagation Neural Network(BPN) and Support Vector Machine (SVM), set up six kinds of model to predict the risk of GERD, In addition, the Case-Based Reasoning(CBR) was used to evaluate whether the patients were at risk of GERD, and an assessment interface was designed to facilitate the clinical risk assessment of complications.

Each prediction model was verified by K-fold interaction and adjusted by the grid search method. The model efficiency was above 85% classification accuracy and the area under the ROC curve was above 0.879. Friedman's test found significant performance differences between the prediction models under the classification accuracy and the area under the ROC curve. Therefore, paired samples T were further used to test the individual strengths and disadvantages of the prediction models. On the basis of classification accuracy measurement, the model performance of GALR combined with BPNN and PSO combined with SVM and PSO combined with BPNN was equally excellent, with the highest accuracy of 95.16%. On the basis of the area measurement under the ROC curve, the performance of the model combining GALR with BPNN and PSO with BPNN, and GALR with SVM is equally excellent, and the maximum area under ROC curve is 0.95803092. The efficiency of the evaluation system was evaluated by K-fold interaction verification. The classification accuracy of each model was above 92.85%, and the area under the ROC curve was above 0.928. With Friedman test being conducted, it is found that three algorithms are equally suitable for being implemented in evaluation models as a medium of computing medical record similarity. Namely, there exist no significant performance differences between evaluation models. The prediction and evaluation models proposed by current study may serve as a reference for medical institutions. Therefore, all of them are suitable for weight calculation of evaluation system, and this study can provide a reference for the prediction and evaluation of related medical institutions.

摘要 ...i
Abstract ...iii
誌謝 ...v
目錄 ...vi
表目錄 ...xi
圖目錄 ...xiii
第一章 緒論 ...1
1.1 研究背景 ...1
1.2 研究動機 ...2
1.3 研究目的 ...3
1.4 研究範圍 ...3
1.5 研究流程 ...3
第二章 文獻探討 ...5
2.1 慢性阻塞性肺病 ...5
2.1.1 慢性阻塞性肺病簡介 ...5
2.1.2 慢性阻塞性肺病的診斷與定義 ...6
2.1.3 影響慢性阻塞性肺病發展與進展的危險因子 ...13
2.1.4 慢性阻塞性肺病患者治療現況與負擔 ...15
2.2 胃食道逆流 ...17
2.2.1 胃食道逆流簡介 ...17
2.2.2 胃食道逆流症狀與診斷 ...18
2.2.3 胃食道逆流改善與治療 ...21
2.3 基因邏輯斯迴歸演算法 ...23
2.3.1 基因演算法介紹 ...23
2.3.2 基因演算法的步驟 ...24
2.3.3 邏輯斯迴歸分析與基因演算法結合 ...26
2.3.4 基因演算法之相關應用 ...27
2.4 粒子群演算法 ...28
2.4.1 粒子群演算法介紹 ...28
2.4.2 粒子群演算法流程 ...29
2.4.3 粒子群演算法之應用 ...31
2.5 交叉熵演算法 ...32
2.5.1 交叉熵演算法之介紹 ...32
2.5.2 交叉熵演算法之架構 ...32
2.5.3 交叉熵演算法相關應用 ...33
2.6 支援向量機 ...34
2.6.1 支援向量機介紹 ...34
2.6.2 支援向量機架構 ...34
2.6.3 支援向量機之應用 ...37
2.7 倒傳遞類神經網路 ...38
2.7.1 倒傳遞類神經網路之介紹 ...38
2.7.2 倒傳遞類神經網路之架構 ...39
2.7.3 倒傳遞類神經網路之應用 ...42
2.8 案例式推理系統 ...43
2.8.1 案例式推理系統之介紹 ...43
2.8.2 案例式推理系統之架構 ...43
2.8.3 案例式推理之優缺點 ...44
2.8.4 案例式推理之應用 ...45
第三章 研究方法 ...47
3.1 研究問題 ...47
3.1.1 研究對象 ...47
3.1.2 研究資料來源 ...47
3.2 研究方法 ...48
3.2.1 研究方法架構 ...48
3.2.2 研究資料前處理 ...51
3.3 定義研究風險因子變數 ...51
3.4 研究風險因子編碼 ...52
3.5 數據準備與模型建構 ...54
3.5.1 研究使用之軟體與數據準備 ...54
3.5.2 預測模型與評估系統之建構 ...54
3.6 演算法參數調整 ...55
3.6.1 粒子群最佳化演算法 ...55
3.6.2 基因邏輯斯迴歸演算法 ...55
3.6.3 交叉熵演算法 ...56
3.6.4 倒傳遞類神經網路 ...56
3.6.5 支援向量機 ...58
3.7 預測模型超參數調校與模型效能評估 ...58
3.7.1 K疊交互驗證 ...58
3.7.2 混淆矩陣 ...60
3.7.3 ROC曲線與ROC曲線下面積 ...63
3.8 模型差異之統計檢定 ...65
第四章 研究結果 ...67
4.1 資料敘述性統計與視覺化 ...67
4.2 PSO實驗分析 ...68
4.2.1 PSO參數設定 ...68
4.2.2 PSO結合SVM實驗分析 ...70
4.2.3 PSO結合BPNN實驗分析 ...72
4.3 GALR實驗分析 ...73
4.3.1 GALR初始參數設定 ...74
4.3.2 GALR權重分析 ...74
4.3.3 GALR結合SVM實驗分析 ...75
4.3.4 GALR結合BPNN實驗分析 ...77
4.4 CE實驗分析 ...78
4.4.1 CE參數設定 ...79
4.4.2 CE權重分析 ...79
4.4.3 CE結合SVM實驗分析 ...80
4.4.4 CE結合BPNN實驗分析 ...82
4.5 預測模型參數設定 ...83
4.5.1 SVM初始參數設定 ...84
4.5.2 BPNN初始參數設定 ...84
4.6 各模型比較 ...85
4.6.1 預測模型實驗結果之比較 ...85
4.6.2 模型差異統計檢定 ...86
4.7 案例式推理系統之建立 ...87
4.7.1 案例式推理系統運作與權重值設定 ...87
4.7.2 胃食道逆流評估系統使用介面 ...88
4.7.3 案例式推理評估系統之系統效能驗證 ...95
第五章 結論 ...97
參考文獻 ...99
Extended Abstract ...113

1.王鈺分、謝旻凌、陳建維 (2021 年 7 月) 慢性阻塞性肺病合併肺動脈高壓及右心衰竭患者使用體外膜氧合器之呼吸照護經驗。呼吸治療， 20(2)，157–158。
2.台灣慢性阻塞性肺病學會 (2012)，台灣慢性阻塞性肺病學會，取自：
http://www.tacopd.org.tw/origin/origin_01.asp 讀取日：2021，12，9。
3.台灣胸腔暨重症加護學會 (2017)，自在呼吸健康網 | 全台灣第一個呼吸道民眾衛教專業網站，取自：
http://www.asthma-copd.tw 讀取日：2022，04，14。
4.吳文傑 (2021 年 8 月) 小心胃食道逆流找上身。消費者報導雜誌， 484，67–70。
5.吳英黛 (2012)，社團法人中華民國物理治療師公會全國聯合會，取自： http://www.pt.org.tw/pt_news_detail.php?Bid=8&Id=187 讀取日：2021，12，9。
6.何瑛曌 (2021 年 7 月) 防疫保「胃」戰，在家如何舒緩胃食道逆流。彰基院訊， 38(7)，10–11。
7.林哲暐、李兆殷、吳宗樺、吳皓瑋、葉家成、楊雅雯、侯家鼎 (2006)，慢性阻塞性肺部疾病醫療資源耗用之研究-以全民健保資料庫為例，取自： https://nhird.nhri.org.tw/talk_05_abs_305 讀取日：2021，12，9。
8.陳家弘 (2018)，慢性阻塞性肺病APP 掌握病情最及時 | 醫療新聞，中國醫藥大學附設醫院，取自：
http://www.cmuh.cmu.edu.tw/NewsInfo/NewsArticle?no=4231 讀取日：2022，01，12。
9.黃家微、嚴麗君、劉嘉恩 (2021 年 7 月)。 一位獨居慢性阻塞性肺病病人之護理經驗。若瑟醫護雜誌， 15(1)，107–117。
10.董宥汝、羅景全 (2017 年 1 月)。 非侵蝕性食道逆流疾病。臨床醫學， 79(1)，32–38。
11.臺北榮民總醫院胸腔部 (2016)，臺北榮民總醫院胸腔部，取自： https://wd.vghtpe.gov.tw/cmd/Fpage.action?muid=3031&fid=6460 讀取日：2021，12，9。
12.衛生福利部國民健康署 (2016)，衛生福利部國民健康署，取自： https://www.hpa.gov.tw/Pages/List.aspx?nodeid=41 讀取日：2021，12，9。
13.戴有志、王薏婷、黃慧君 (2009 年 12 月 1 日)。 慢性阻塞性肺病。北台灣中醫醫學雜誌， 1(2)，78–¬¬86。
14.嚴助成、張維國、謝財源 (2005 年 12 月)。 胃食道逆流疾病的認識與治療。臺灣臨床藥學雜誌， 13(2)，55–62。
15.Ajmera, M., Raval, A. D., Shen, C., & Sambamoorthi, U. (2014). Explaining the increased health care expenditures associated with gastroesophageal reflux disease among elderly Medicare beneficiaries with chronic obstructive pulmonary disease: A cost-decomposition analysis. International Journal of Chronic Obstructive Pulmonary Disease, 9.
16.Antunes C., Aleem A., & Curtis S. A. (2021). Gastroesophageal Reflux Disease. StatPearls [Internet]. StatPearls Publishing, available at
https://www.ncbi.nlm.nih.gov/books/NBK441938/ retrieved Frbruary 26, 2022.
17.Ardelean, S. M., & Udrescu, M. (2022). Graph coloring using the reduced quantum genetic algorithm. Peer Journal Computer Science.
18.Boeree, M. J., Peters, F. T. M., Postma, D. S., & Kleibeuker, J. H. (1998). No effects of high-dose omeprazole in patients with severe airway hyperresponsiveness and (a)symptomatic gastro-oesophageal reflux. European Respiratory Journal, 11(5), 1070–1074.
19.Bor, S., Kitapcioglu, G., Solak, Z. A., Ertilav, M., & Erdinc, M. (2010). Prevalence of gastroesophageal reflux disease in patients with asthma and chronic obstructive pulmonary disease. Journal of Gastroenterology and Hepatology, 25(2), 309–313.
20.Casanova, C., Baudet, J. S., del Valle Velasco, M., Martin, J. M., Aguirre-Jaime, A., Pablo de Torres, J., & Celli, B. R. . (2004). Increased gastro-oesophageal reflux disease in patients with severe COPD. European Respiratory Journal, 23(6), 841–845.
21.Chen, Y. (2021). Business English Translation Model Based on BP Neural Network Optimized by Genetic Algorithm. Computational Intelligence and Neuroscience.
22.Cleveland Clinic. (2019). GERD (Chronic Acid Reflux): Symptoms, Treatment, & Causes. Cleveland Clinic, available at
https://my.clevelandclinic.org/health/diseases/17019-gerd-or-acid-reflux-or-heartburn-overview retrieved Frbruary 26, 2022.
23.Felson S. (2021). COPD (Chronic Obstructive Pulmonary Disease). WebMD, available at
https://www.webmd.com/lung/copd/10-faqs-about-living-with-copd retrieved January 18,2022.
24.Guerrero, J. I., Miró-Amarante, G., & Martín, A. (2022). Decision support system in health care building design based on case-based reasoning and reinforcement learning. Expert Systems with Applications, 187.
25.Han, M.-Z., Hsiue, T.-R., Tsai, S.-H., Huang, T.-H., Liao, X.-M., & Chen, C.-Z. (2018). Validation of the GOLD 2017 and new 16 subgroups (1A–4D) classifications in predicting exacerbation and mortality in COPD patients. International Journal of Chronic Obstructive Pulmonary Disease, Volume 13, 3425–3433.
26.Hillas, G., Perlikos, F., Tsiligianni, I., & Tzanakis, N. (2015). Managing comorbidities in COPD. International Journal of Chronic Obstructive Pulmonary Disease, 10, 95–109.
27.Huang, C., Liu, Y., & Shi, G. (2020). A systematic review with meta-analysis of gastroesophageal reflux disease and exacerbations of chronic obstructive pulmonary disease. BMC Pulmonary Medicine, 20, 2.
28.Iwakiri, K., Fujiwara, Y., Manabe, N., Ihara, E., Kuribayashi, S., Akiyama, J., Kondo, T.,et al. (2022). Evidence-based clinical practice guidelines for gastroesophageal reflux disease 2021. Journal of Gastroenterology, 57(4), 267–285.
29.Khosravanian, R., & Aadnøy, B. S. (2022). Chapter Twelve—Case-based reasoning (CBR) in digital well planning and construction. R. Khosravanian & B. S. Aadnøy, Methods for Petroleum Well Optimization pp.477–521. Gulf Professional Publishing , available at
https://www.sciencedirect.com/science/article/pii/B9780323902311000030
retrieved April 13,2022.
30.Kostadinov S. (2019). Understanding Backpropagation Algorithm. Medium, available at
https://towardsdatascience.com/understanding-backpropagation-algorithm-7bb3aa2f95fd retrieved March 8,2022.
31.Lambora, A., Gupta, K., & Chopra, K. (2019). Genetic Algorithm- A Literature Review. 2019 International Conference on Machine Learning, Big Data, Cloud and Parallel Computing (COMITCon),pp.380–384.
32.Lee, J., Jung, H. M., Kim, S. K., Yoo, K. H., Jung, K.-S., Lee, S. H., & Rhee, C. K. (2019). Factors associated with chronic obstructive pulmonary disease exacerbation, based on big data analysis. Scientific Reports.
33.Liu, T., & Zou, G.(2021). Evaluation of Mechanical Properties of Materials Based on Genetic Algorithm Optimizing BP Neural Network. Computational Intelligence and Neuroscience, 2021.
34.Mokhlesi, B., Morris, A. L., Huang, C.-F., Curcio, A. J., Barrett, T. A., & Kamp, D. W. (2001). Increased Prevalence of Gastroesophageal Reflux Symptoms in Patients With COPD. Chest, 119(4), 1043–1048.
35.Moradi, S., Tapak, L., & Afshar, S. (2022). Identification of Novel Noninvasive Diagnostics Biomarkers in the Parkinson’s Diseases and Improving the Disease Classification Using Support Vector Machine. BioMed Research Internationa.
36.National Heart, Lung, and Blood Institute. (2022). COPD. Text, National Library of Medicine, available at
https://medlineplus.gov/copd.html retrieved January 10,2022.
37.Niemeyer, J. F., Rudolf, S., Kvaratskhelia, L., Mennenga, M., & Herrmann, C. (2022). A creativity-driven Case-Based Reasoning Approach for the systematic Engineering of Sustainable Business Models. Procedia CIRP, The 29th CIRP Conference on Life Cycle Engineering, April 4 – 6, 2022, Leuven, Belgium., 105, 470–475.
38.Nonlinear Support Vector Machine—An overview | ScienceDirect Topics. (2022), available at
https://www.sciencedirect.com/topics/engineering/nonlinear-support-vector-machine
retrieved April 16,2022
39.Pace, F., Santilano, A., & Godio, A. (2021). A Review of Geophysical Modeling Based on Particle Swarm Optimization. Surveys in Geophysics, 42(3), 505–549.
40.Piotrowski, A. P., & Piotrowska, A. E. (2021). Differential evolution and particle swarm optimization against COVID-19. Artificial Intelligence Review, pp.1–71.
41.Popescu, V.-B., Kanhaiya, K., Năstac, D. I., Czeizler, E., & Petre, I. (2022). Network controllability solutions for computational drug repurposing using genetic algorithms. Scientific Reports.
42.Putcha, N., Drummond, M. B., Wise, R. A., & Hansel, N. N. (2015). Comorbidities and Chronic Obstructive Pulmonary Disease: Prevalence, Influence on Outcomes, and Management. Seminars in respiratory and critical care medicine, 36(4), 575–591.
43.Quan, Z., Yan, G., Wang, Z., Li, Y., Zhang, J., Yang, T., & Piao, H. (2021). Current status and preventive strategies of chronic obstructive pulmonary disease in China: A literature review. Journal of Thoracic Disease, 13(6), 3865–3877.
44.Riesco Miranda, J. A., Marca-Frances, G., & Jimenez-Ruiz, C. A. (2018). Perception and Awareness of Chronic Obstructive Pulmonary Disease,
Chronic Bronchitis and Pulmonary Emphysema in the Spanish Urban Population. Archivos de Bronconeumología (English Edition), 54(6), 352–353.
45.Salman, I. (2018). Impact of Metaheuristic Iteration on Artificial Neural Network Structure in Medical Data. Processes, 6, available at
https://www.mdpi.com/2227-9717/6/5/57 retrieved April 16,2022
46.Spindelböck, T., Ranftl, S., & von der Linden, W. (2021). Cross-Entropy Learning for Aortic Pathology Classification of Artificial Multi-Sensor Impedance Cardiography Signals. Entropy, 23(12), 1661.
47.Spirometry.guru. (2022). Reversibility, available at
https://www.spirometry.guru/reversibility.html retrieved April 15,2022
48.Sweis, R., & Fox, M. (2020). The global burden of gastro-oesophageal reflux disease: More than just heartburn and regurgitation. The Lancet Gastroenterology & Hepatology, 5(6), 519–521.
49.Takada, K., Matsumoto, S., Kojima, E., Iwata, S., Okachi, S., Ninomiya, K., Morioka, H.et al. (2011). Prospective evaluation of the relationship between acute exacerbations of COPD and gastroesophageal reflux disease diagnosed by questionnaire. Respiratory Medicine, 105(10), 1531–1536.
50.Tapak, L., Afshar, S., Afrasiabi, M., Ghasemi, M. K., & Alirezaei, P. (2021). Application of Genetic Algorithm-Based Support Vector Machine in Identification of Gene Expression Signatures for Psoriasis Classification: A Hybrid Model. BioMed Research Internationa, 2021.
51.Tsai, C.-L., Lin, Y.-H., Wang, M.-T., Chien, L.-N., Jeng, C., Chian, C.-F., Perng, W.-C.et al. (2015). Gastro-oesophageal reflux disease increases the risk of intensive care unit admittance and mechanical ventilation use among patients with chronic obstructive pulmonary disease: A nationwide population-based cohort study. Critical Care, 19(1), 110.
52.Wang, K., Yang, Y., Reniers, G., Li, J., & Huang, Q. (2022). Predicting the spatial distribution of direct economic losses from typhoon storm surge disasters using case-based reasoning. International Journal of Disaster Risk Reduction.
53.Wang, K.-F., An, J., Wei, Z., Cui, C., Ma, X.-H., Ma, C., & Bao, H.-Q. (2022). Deep Learning-Based Imbalanced Classification With Fuzzy Support Vector Machine. Frontiers in Bioengineering and Biotechnology.
54.Wang, Y.-J., Lang, X.-Q., Wu, D., He, Y.-Q., Lan, C.-H., Xiao-Xiao, Wang, B.et al. (2020). Salivary Pepsin as an Intrinsic Marker for Diagnosis of Sub-types of Gastroesophageal Reflux Disease and Gastroesophageal Reflux Disease-related Disorders. Journal of Neurogastroenterology and Motility, 26(1), 74–84.
55.World Health Organization (2020a), NCD Country Capacity Survey, available at https://www.who.int/teams/ncds/surveillance/monitoring-capacity/ncdccs
retrieved December 09, 2021
56.World Health Organization (2020b).The top 10 causes of death, available at https://www.who.int/news-room/fact-sheets/detail/the-top-10-causes-of-death
retrieved December 09, 2021
57.World Health Organization (2021).Chronic obstructive pulmonary disease (COPD), available at
https://www.who.int/news-room/fact-sheets/detail/chronic-obstructive-pulmonary-disease-(copd) retrieved December 09, 2021
58.Wu, M.-T. (2022). Confusion matrix and minimum cross-entropy metrics based motion recognition system in the classroom. Scientific Reports, 12, 3095.
59.Xie, Z., Huang, X., & Liu, W. (2022). Subpopulation Particle Swarm Optimization with a Hybrid Mutation Strategy. Computational Intelligence and Neuroscience, 2022.
60.Yeung, M., Sala, E., Schönlieb, C.-B., & Rundo, L. (2022). Unified Focal loss: Generalising Dice and cross entropy-based losses to handle class imbalanced medical image segmentation. Computerized Medical Imaging and Graphics, 95.
61.Yu, H., & Wang, T. (2021). A Method for Real-Time Fault Detection of Liquid Rocket Engine Based on Adaptive Genetic Algorithm Optimizing Back Propagation Neural Network. Sensors (Basel, Switzerland), 21(15), 5026.
62.Zhu, W., Song, Y., & Xiao, Y. (2020). Support vector machine classifier with huberized pinball loss. Engineering Applications of Artificial Intelligence, 36(5), 984-997.