簡易檢索 / 詳目顯示

回結果列表
研究生： 呂易儒
Yi-Ru Lu
論文名稱： 神經網路於敗血病特徵及生理訊號異常偵測
Sepsis Characteristic and Biosignal Anomaly Detection with A Neural Network Approach
指導教授： 張以全
Peter I-Tsyuen Chang
口試委員: 張以全
Peter I-Tsyuen Chang
張春梵
Chun-Fan Chang
許昕
Hsin Hsiu
學位類別： 碩士
Master
系所名稱： 工程學院 - 機械工程系
Department of Mechanical Engineering
論文出版年： 2021
畢業學年度： 110
語文別： 英文
論文頁數： 75
中文關鍵詞： 敗血病 、生物訊號 、訊號解析 、變模態分解 、神經網路
外文關鍵詞： Sepsis, Biosignal, Signal decomposition, Variational mode decomposi­ tion(VMD), Deep neural neetwork
相關次數： 點閱：243  下載：0 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 敗血病在加護病房是常出現且容易造成死亡的病症，而通常加護病房 (ICU) 的維持費用以及使用的醫療資源在一家醫院裡也佔了大宗，因此如果越早偵測到 敗血病的發作，點出危急的病人，就能夠讓醫生更快的下達醫療指示並給予治療 方向。所以在本論文中，希望使用先前已經搜集到的豬隻生理訊號資料，訓練一 個敗血病預測模型。我們利用全身性炎症反應症候群(SIRS)作為敗血病的評分 標準，並透過數學方法與機器學習分類器結合，辨識敗血症的相關特徵，捨棄信 號中許多無關的雜訊，整合成一種預測模型。
    本論文開發一種透過豬隻資料並利用人工神經網路去偵測敗血病的方法。其 中使用變模態分解(VMD)的方法來擷取這些生理訊號的相關特徵，並作為人工 神經網路分類器的訓練輸入，最後希望此模型可納入自動化醫學量測，未來可在 臨床研究中使用，作為一個醫療診斷輔助的工具，在即將病發之前及早偵測，作 為臨床醫師監測與診斷之重要參考，加快決定治療時機。

    Sepsis is the common disease that causes death and occurs in the Intensive Care Unit(ICU). Usually, the maintenance cost and the medical resources used in the Inten­ sive Care Unit(ICU) both occupy a large amount in a hospital. Therefore, as fast as the detection of sepsis onset and point out the critical patient, the doctor can give medical instructions and give the way for treatment immediately. In this thesis, we want to use the pig’s bio­signal to train a sepsis prediction model. We use the systemic inflamma­ tory response syndrome(SIRS) as a scoring standard and use mathematical method and combine with machine learning classifier to identify the characteristics of sepsis, remove many unrelated features and integrate into a predictive model.
    This research is expected to develop a method of detecting sepsis through pig’s data by using artificial neural networks. By using Variational Mode Decomposition(VMD) that can capture the related feature of these bio­signals, then these features can be training input of artificial neural network classifier. In the future, we hope this model can be included in the automatic medical measurement and can be applied in clinical research as a medical decision support tool. It can diagnosis before the disease occurs and become an important reference for clinical doctor or diagnosis then speeding up the timing of treatment.

    Abstract in Chinese.................................. I Abstract in English .................................. II Acknowledgements.................................. III Contents........................................ IV List of Figures..................................... VII List of Tables ..................................... IX 1 Introduction.................................... 1 1.1 Motivation.................................. 1 1.2 Researchobjectives............................. 2 1.3 Thesisoutline................................ 3 2 LiteratureReview ................................. 4 2.1 DefinitionofSepsis............................. 4 2.1.1 Correlation of SIRS and qSOFA in the emergency department . . 4 2.1.2 CorrelationofSIRSandqSOFAinICU . . . . . . . . . . . . . . 5 2.1.3 Sensitivity and Specificity of SIRS and qSOFA . . . . . . . . . . 5 2.2 Sepsisdetectionandprediction....................... 8 2.3 MedicaldataandResearchDataset..................... 11 2.4 Artificialneuralnetwork .......................... 11 2.4.1 Machinelearningtoolbox ..................... 13 2.5 Summary .................................. 13 3 Datapreparation.................................. 14 3.1 Datacollectionandcriteria......................... 14 3.1.1 Inclusioncriteria .......................... 14 3.1.2 Lipopolysaccharide(LPS) ..................... 15 3.2 Rawdatapreliminaryclassification..................... 16 3.2.1 Missingdataprocessing ...................... 16 3.2.2 Casegroupandcontrolgroup ................... 17 3.3 Featureselection .............................. 18 3.4 Featureextraction.............................. 21 3.4.1 Dataalignment ........................... 26 3.4.2 Empiricalmodedecomposition .................. 27 3.4.3 SelectingtheIMFs ......................... 29 3.4.4 VariationalModeDecomposition ................. 30 3.4.5 Reconstructsignal ......................... 35 3.5 Slidingwindow ............................... 37 3.6 Labelingprocess .............................. 38 3.7 Finaldataset................................. 38 4 Modeldevelopment................................ 40 4.1 Artificialneuralnetwork .......................... 40 4.1.1 Feedforwardnetwork........................ 40 4.1.2 Artificial neurons.......................... 41 4.1.3 Backpropagation.......................... 44 4.1.4 TrainingAlgorithm......................... 44 4.2 Trainingdataseparation........................... 47 4.3 Performancemetrics ............................ 48 4.3.1 Confusionmatrix.......................... 48 4.3.2 ROCCurveandAreaUndertheCurve . . . . . . . . . . . . . . 50 5 Resultsanddiscussion............................... 52 5.1 Adjustthemodel .............................. 52 5.2 Testnewdataforthemodel......................... 55 5.3 Discussion.................................. 56 5.4 Limitations ................................. 59 5.5 ConclusionandFuturework ........................ 59 References....................................... 61

    [1] Carolin Fleischmann, Daniel O Thomas­Rueddel, Michael Hartmann, Christiane S Hartog, Tobias Welte, Steffen Heublein, Ulf Dennler, and Konrad Reinhart. Hospital incidence and mortality rates of sepsis. Deutsches Arzteblatt international, 113(10):159–166, 03 2016.
    [2] Derek C. Angus, Walter T. Linde­Zwirble, Jeffrey Lidicker, Gilles Clermont, Joseph Carcillo, and Michael R. Pinsky. Epidemiology of severe sepsis in the united states: Analysis of incidence, outcome, and associated costs of care. Critical Care Medicine, 29(7), 2001.
    [3] Yen­Jung Chen, Fu­Lun Chen, Jin­Hua Chen, Man­Tzu Marcie Wu, Yen­Ling Chen, Du­Shieng Chien, and Yu Ko. Epidemiology of sepsis in taiwan. Medicine (Baltimore), vol.98(no.20), May 2019.
    [4] Ministry of Health and Welfare. Statistics on causes of death in 2019, 06 2019.
    [5] Mervyn Singer, Clifford S Deutschman, Christopher Warren Seymour, Manu Shankar­Hari, Djillali Annane, Michael Bauer, Rinaldo Bellomo, Gordon R Bernard, Jean­Daniel Chiche, Craig M Coop­ ersmith, Richard S Hotchkiss, Mitchell M Levy, John C Marshall, Greg S Martin, Steven M Opal, Gordon D Rubenfeld, Tom van der Poll, Jean­Louis Vincent, and Derek C Angus. The third inter­ national consensus definitions for sepsis and septic shock (sepsis­3). JAMA, 315(8):801–810, Feb 2016.
    [6] RogerC.Bone,RobertA.Balk,FrankB.Cerra,R.PhillipDellinger,AlanM.Fein,WilliamA.Knaus, Roland M. H. Schein, and William J. Sibbald. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Chest, 101(6):1644–1655, 1992.
    [7] J.L.Vincent,R.Moreno,J.Takala,S.Willatts,A.DeMendonça,H.Bruining,C.K.Reinhart,P.M. Suter, and L. G. Thijs. The sofa (sepsis­related organ failure assessment) score to describe organ dysfunction/failure. Intensive Care Medicine, 22(7):707–710, 1996.
    [8] Julian M. Williams, Jaimi H. Greenslade, Juliet V. McKenzie, Kevin Chu, Anthony F. T. Brown, and Jeffrey Lipman. Systemic inflammatory response syndrome, quick sequential organ function assessment, and organ dysfunction: Insights from a prospective database of ed patients with infection. Chest, 151(3):586–596, 2017.
    [9] Yonathan Freund, Najla Lemachatti, Evguenia Krastinova, Marie Van Laer, Yann­Erick Claessens, Aurélie Avondo, Céline Occelli, Anne­Laure Feral­Pierssens, Jennifer Truchot, Mar Ortega, Bruno Carneiro, Julie Pernet, Pierre­Géraud Claret, Fabrice Dami, Ben Bloom, Bruno Riou, Sébastien Beaune, and for the French Society of Emergency Medicine Collaborators Group. Prognostic Ac­ curacy of Sepsis­3 Criteria for In­Hospital Mortality Among Patients With Suspected Infection Pre­ senting to the Emergency Department. JAMA, 317(3):301–308, 01 2017.
    61
    [10] Eamon P. Raith, Andrew A. Udy, Michael Bailey, Steven McGloughlin, Christopher MacIsaac, Ri­ naldo Bellomo, David V. Pilcher, for the Australian, New Zealand Intensive Care Society (ANZ­ ICS) Centre for Outcomes, and Resource Evaluation (CORE). Prognostic Accuracy of the SOFA Score, SIRS Criteria, and qSOFA Score for In­Hospital Mortality Among Adults With Suspected In­ fection Admitted to the Intensive Care Unit. JAMA, 317(3):290–300, 01 2017.
    [11] Åsa Askim, Florentin Moser, Lise T. Gustad, Helga Stene, Maren Gundersen, Bjørn Olav Åsvold, Jostein Dale, Lars Petter Bjørnsen, Jan Kristian Damås, and Erik Solligård. Poor performance of quick­sofa (qsofa) score in predicting severe sepsis and mortality –a prospective study of patients admitted with infection to the emergency department. Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine, 25(1):56, 2017.
    [12] Jean­LouisVincent,GregSMartin,andMitchellMLevy.qsofadoesnotreplacesirsinthedefinition of sepsis. Crit Care, 20(1):210, Jul 2016.
    [13] MatthieuScherpf,FelixGräßer,HagenMalberg,andSebastianZaunseder.Predictingsepsiswithare­ current neural network using the mimic iii database. Computers in Biology and Medicine, 113:103395, 2019.
    [14] HyeJinKamandHaYoungKim.Learningrepresentationsfortheearlydetectionofsepsiswithdeep neural networks. Computers in Biology and Medicine, 89:248–255, 2017.
    [15] Katharine Henry, David Hager, Peter Pronovost, and Suchi Saria. A targeted real­time early warning score (trewscore) for septic shock. Science translational medicine, 7:299ra122, 08 2015.
    [16] QingqingMao,MelissaJay,JanaLHoffman,JacobCalvert,ChristopherBarton,DavidShimabukuro, Lisa Shieh, Uli Chettipally, Grant Fletcher, Yaniv Kerem, Yifan Zhou, and Ritankar Das. Multicentre validation of a sepsis prediction algorithm using only vital sign data in the emergency department, general ward and icu. BMJ Open, 8(1):e017833, 01 2018.
    [17] Jacob S. Calvert, Daniel A. Price, Uli K. Chettipally, Christopher W. Barton, Mitchell D. Feldman, Jana L. Hoffman, Melissa Jay, and Ritankar Das. A computational approach to early sepsis detection. Computers in Biology and Medicine, 74:69–73, 2016.
    [18] ChristopherBarton,UliChettipally,YifanZhou,ZiruiJiang,AnnaLynn­Palevsky,SidneyLe,Jacob Calvert, and Ritankar Das. Evaluation of a machine learning algorithm for up to 48­hour advance prediction of sepsis using six vital signs. Computers in Biology and Medicine, 109, 04 2019.
    [19] Thomas Desautels, Jacob Calvert, Jana Hoffman, Melissa Jay, Yaniv Kerem, Lisa Shieh, David Shimabukuro, Uli Chettipally, Mitchell Feldman, Chris Barton, David Wales, and Ritankar Das. Pre­ diction of sepsis in the intensive care unit with minimal electronic health record data: A machine learning approach. JMIR Medical Informatics, 4, 09 2016.
    [20] Takenori Kato, Mohamed Hussein, Takahiro Sugiura, Satoshi Suzuki, Sumio Fukuda, Tanaka Tai­ hei, Ineko Kato, and Hajime Togari. Development and characterization of a novel porcine model of neonatal sepsis. Shock (Augusta, Ga.), 21:329–35, 05 2004.
    62

    [21] Antoine Guillon, Sebastien Preau, Jérôme Aboab, Eric Azabou, Boris Jung, Stein Silva, Julien Tex­ toris, Fabrice Uhel, Dominique Vodovar, Lara Zafrani, Nicolas de Prost, and Peter Radermacher. Pre­ clinical septic shock research: why we need an animal icu. Ann Intensive Care, 9(1):66, Jun 2019.
    [22] C. J. Schuh. Sepsis and septic shock analysis using neural networks. In NAFIPS 2007 ­ 2007 Annual Meeting of the North American Fuzzy Information Processing Society, pages 650–654, 2007.
    [23] FilippoAmato,AlbertoLópez­Rodríguez,EladiaPeña­Méndez,PetrVaňhara,AlešHampl,andJosef Havel. Artificial neural networks in medical diagnosis. J Appl Biomed, 11:47–58, 12 2013.
    [24] Stephen R Wisniewski, Andrew C Leon, Michael W Otto, and Madhukar H Trivedi. Prevention of missing data in clinical research studies. Biol Psychiatry, 59(11):997–1000, Jun 2006.
    [25] Hyun Kang. The prevention and handling of the missing data. Korean J Anesthesiol, 64(5):402–406, May 2013.
    [26] RobertMHamerandPippaMSimpson.Lastobservationcarriedforwardversusmixedmodelsinthe analysis of psychiatric clinical trials. Am J Psychiatry, 166(6):639–641, Jun 2009.
    [27] LucasMFleuren,ThomasLTKlausch,CharlotteLZwager,LindaJSchoonmade,TingjieGuo,LucaF Roggeveen, Eleonora L Swart, Armand RJ Girbes, Patrick Thoral, Ari Ercole, et al. Machine learning for the prediction of sepsis, a systematic review and meta­analysis of diagnostic test accuracy. Intensive Care Medicine, 46(3):383–400, 2020.
    [28] J. D. Young. The heart and circulation in severe sepsis. BJA: British Journal of Anaesthesia, 93(1):114–120, 07 2004.
    [29] Norden E. Huang, Zheng Shen, Steven R. Long, Manli C. Wu, Hsing H. Shih, Quanan Zheng, Nai­ Chyuan Yen, Chi Chao Tung, and Henry H. Liu. The empirical mode decomposition and the hilbert spectrum for nonlinear and non­stationary time series analysis. 454(1971):903–995, 2021/06/15/ 1998.
    [30] Konstantin Dragomiretskiy and Dominique Zosso. Variational mode decomposition. IEEE Transac­ tions on Signal Processing, 62(3):531–544, 2014.
    [31] M.T. Hagan and M.B. Menhaj. Training feedforward networks with the marquardt algorithm. IEEE Transactions on Neural Networks, 5(6):989–993, 1994.
    [32] Bogdan M. Wilamowski Hao Yu. Levenberg–marquardt training. The Industrial Electronics Hand­ book, 2011.
    [33] Sam Roweis. Levenberg­marquardt optimization. Note, University Of Toronto, 1996.

    QR CODE