臺灣博碩士論文加值系統

本研究是以心臟血管疾病為範圍，針對資料探勘中的分類問題做為研究主軸，希望以探索性資料分析技術，發展出一套合理與有效的臨床輔助預知模型(Prognosis model)。方法為：使用子集合屬性選擇(subset attribute selection)演算法做為特徵選取，以基本模型(base model)方法與集成方法(ensemble methods)的計算結果做比較。當中，基礎模型，選擇以人工智慧類神經網路(Artificial Neural Network)、支援向量機(Support Vector Machine)、決策樹(Decision Tree)等方法為主。接著以組合異質模型的方式，來建立心臟血管疾病術後的預測模型。研究結果為：發展出之模型在準確度，或者是AUC(Area under the Receiver Operating Characteristic curve)上，有顯著之效能改善。結論為：本研究所建構出的預知模型，用於心臟血管疾病診斷來說，具有良好自動化之預期效能。其可應用於決策支援系統，用於提供外科醫師作為診斷預測術後的併發症(Como3)、術後病人在加護病房的時間(T_ICU)、術後住院時間(Length of hospital stay)、術後第1次心臟超音波檢查的左心室功能(Echo 2 Lvef)、術後心臟的分類等級(Alive_Fc)的參考。

This study focuses on using exploratory data analysis to develop cardiovsacular disease prognosis models for making clinical decisions. Method: used the subset attribute selection as the feature selection algorithm, and compared the prognosis models performance by using base model methods and ensemble methods. For the basic models, we choose Artificial Network Neural intelligence (ANN), Support Vector Machine (SVM), and Decision Tree (DT) as the main test models; followed by using ensemble method to build a prediction model for post-operative cardiovascular surgery. The results have shown significant improvements in the accuracy of the prediction model, and in the AUC (Area under the Receiver Operating Characteristic curve). Conclusion: the post-operative prediction model for cardiovascular disease diagnosis constructed by our study has shown satisfactory prediction ability. It can be used in decision analysis system to assist surgeons in predicting Como3, T_ICU, length of hospital stay, Echo 2 Lvef and Alive_Fc.

圖目錄 VI
表目錄 VII
第一章 緒論 1
第一節　研究背景 1
第二節　研究動機 1
第三節　研究目的 2
第四節 研究步驟流程 3
第五節　論文架構 5
第二章 文獻探討 6
第一節 心臟血管相關知識 6
第二節 在醫療資訊領域資料分析方法 10
第三節 機械學習變數選擇技術 18
第四節 機器學習預測模型選擇 23
第五節 集成方法 27
第六節 分類器效能可靠評估方法 29
第七節 資料離散化 34
第三章 研究方法 35
第一節 探索性資料分析方法論 35
第二節 探索性資料分析方法之步驟描述 37
第三節 預測模型建立 39
第四節 資料探勘工具 40
第四章 實證評估 41
第一節 探索性資料分析方法之步驟實做 41
第五章 結論與討論 78
第一節 結論 78
第二節 討論 80
參考文獻 81
附錄一 原始資料庫欄位 85

1.台大醫院院長陳明豐暨台大醫院心臟衰竭中心醫療團隊, 守護一顆心-心臟衰竭診治照護全書. 2009: 原水文化.
2.Purwanto, et al., Prediction Models for Early Risk Detection of Cardiovascular Event. Journal of Medical Systems, 2010.
3.蕭明熙．蔡敬民．許惠恒．陳志鴻．劉秉彥．鄭金寶, 血脂異常關鍵報告. 2007: 原水文化.
4.顏裕庭．許素貞等人, 常見心臟病手冊. 2004: 合記圖書出版社.
5.亞東紀念醫院心臟血管外科衛教園地. Available from: http://depart.femh.org.tw/cvs/06/asp-IN-01.asp?prono=81.
6.Tukey and J. Wilder, Exploratory Data Analysis. 1977: Addison-Wesley.
7.Methods, N.S.e.-H.o.S. Exploratory Data Analysis. Available from: http://www.itl.nist.gov/div898.
8.Hall, M.A. and G. Holmes, Benchmarking attribute selection techniques for discrete class data mining. IEEE Trans Knowledge Data Eng, 2003. 15: p. 1-16.
9.Guyon, I.s. and A. Elisseeff, An Introduction to variable and feature selection. Mach Learning Res, 2003. 3: p. 1157-82.
10.Fiol, G.D. and P.J. Haug, Classification models for the prediction of clinicians’ information needs. Journal of Biomedical Informatics, 2009. 42(1): p. 82-89.
11.劉楊, 特徵選擇方法在信用評估指標選取中應用. 數理統計與管理, 2006. 25(6): p. 667-674.
12.Kira, K. and L.A. Rendell, A practical approach to feature selection. Mach Learning 1992: p. 249-56.
13.Hall, m.A., Correlation-based feature selection for discrete and numeric class machine learning. Proc 17th Int Conf Mach Learning, 2000: p. 359-66.
14.Kohavi, R. and G.H. John, Wrappers for feature subset selection. Artificial Intell, 1997. 97: p. 273-324.
15.王保進, 多變量分析：套裝程式與資料分析。. 2004, 台北市: 高等教育文化.
16.陳鴻烈、蔡大偉, 以主成分分析法探討水庫優養化之動力研究. 水土保持學報, 2009. 40(2): p. 137-162.
17.吳美琴, 應用資料包絡法探討田口方法多品質特性問題之研究. 2007, 逢甲大學工業工程與系統管理研究所.
18.Liu, H. and R. Setiono, A probabilistic approach to feature selection: a filter solution a filter solution. Proc 13th Int conf mach learning, 1996: p. 319-27.
19.洪永祥、江柏儒, 應用IG特徵選取改善SVM多類別分類績效, 2009 第17屆模糊理論及其應用研討會.
20.張斐章、張麗秋、黃浩倫, 類神經網路理論與實務. 2003: 東華書局.
21.葉怡成, 類神經網路模式應用與實作. 2009: 儒林圖書有限公司.
22.Çomak, E. and A. Arslan, A Biomedical Decision Support System Using LS-SVM Classifier with an Efficient and New Parameter Regularization Procedure for Diagnosis of Heart Valve Diseases. Journal of Medical Systems, 2010.
23.盧瑜芬, 使用三種資料探勘演算法-類神經網路、邏輯斯迴歸及決策樹-預測乳患者存活情形之效能比較,國防醫學院公共衛生學研究所流行病學組碩士論文. 2006.
24.Lee, S., Using data envelopment analysis and decision trees for efficiency analysis and recommendation of B2C controls. Decision Support Systems, 2010. 49(4): p. 486-497.
25.Witten, I.H. and E. Frank, Data Mining. 2005: Diane Cerra.
26.Polikar, R., Ensemble based systems in decision making. IEEE Circuits and Systems Magazine, 2006. 6(3): p. 21-45.
27.Zhou and Z. H, Ensemble, ed. I.L. Liu and T. Özsu. 2009, Berlin: Springer: Encyclopedia of database systems.
28.Wang, G., et al., A comparative assessment of ensemble learning for credit scoring. Expert Systems with Applications, 2011. 38(1): p. 223-230.
29.Hansen, L.K. and P. Salamon, Neural network ensembles. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1990. 12(10): p. 993-1001.
30.Dietterich, T.G., Machine learning research: Four current directions. AI Magazine, 1997. 18(4): p. 97-136.
31.Windeatt, T. and G. Ardeshir, Decision tree simplification for classifier ensembles. International Journal of Pattern Recognition, 2004. 18(5): p. 749-776.
32.Friedman, J., T. Hastie, and R. Tibshirani, Additive logistic regression: a statistical view of boosting. The Annals of Statistics, 2000. 28(2): p. 37.
33.Breiman, L., Bagging Predictors. Machine Learning, 1996. 24: p. 123-140.
34.Zhu, D., A hybrid approach for efficient ensembles. Decision Support Systems, 2010. 48(3): p. 480-487.
35.Webb, G.I., MultiBoosting: a technique for combining boosting and wagging. Machine Learning, 2000. 40(2): p. 159-96.
36.Wolpert, D.H., Stacked generalization. Neural Networks 5 1992. 2: p. 241-259.
37.Seewald, A.K., How to make stacking better and faster while also taking care of an unknown weakness. Proc 19th Int Conf Mach Learning, 2002. 554-61.
38.Caruana R, N.-M.A., An empirical comparison of supervised learning
algorithms. Proc 23rd Int Conf Mach Learning, 2006. 148: p. 161-8.
39.Chan, P.K. and S.J. Stolfo, On the accuracy of meta-learning for scalable data mining. JIntell Rev, 1997. 8(1): p. 5-28.
40.林明傑、董子毅, 危險評估中ROC曲線在預測2*2表上與每感度及特異度之關係, 亞洲暴力與性侵害期刊. 2008. p. 64-74.
41.Osareh, A., et al., Comparative exudate classification using support vector machines and neural networks. Springer LNCS 2489, 2002: p. 413-420.
42.Centor, R.M., The use of ROC curves and their analysis. Medical decision making, 1991(11): p. 102-106.
43.李文宗.台大公共衛生院流行病學所副教授. 疾病篩檢診斷工具之評估：從ROC 曲線到Lorenz 曲線到Kullback-Leibler 距離. Available from: http://docs.google.com/viewer?a=v&q=cache:kSyhVVQCy2wJ:aao.sinica.edu.tw/download/jriaaward/34.pdf+Area+under+the+Receiver+Operating+Characteristic+curve&hl=zh-TW&gl=tw&pid=bl&srcid=ADGEEShysk_hHfAK-Sps2WgI27WojNjPeHoGaj7H-F6IhcxBTjKKvHuvU_fqO-mvkdiTURJd4Iw_7XNzrFs2yKOiJ97rcxjqGKBSF5KjIadZfdlnBiF9Wt0sGlmeJ6hkyvMSEzsM_35J&sig=AHIEtbSXYUz00FS6yb_tMVk_fS1b1HEyzA.
44.TechNet, M. 增益圖. Available from: http://technet.microsoft.com/zh-tw/library/ms174947.aspx.
45.Tan, P.-n., M. Steinbach, and V. Kumar, Introduction to Data Mining 2004, Boston: Addison Wesley.
46.韓歆儀, 應用兩階段分類法提升SVM法之分類準確率,成功大學工業與資訊管理研究所碩士論文. 2004.
47.kretschmann, E. and R. Apweiler. Automatic Rule Generation for Protein Annotation with the C4.5 Data-Mining Algorithm Applied on Peptides in Ensembl. in bioinformatics. 2001.
48.Ishikawa, K., Guide To Quality Control. 1986: Infotech Standards.
49.Hsieh, N.-C., Intelligent Postoperative Morbidity Prediction of Heart Disease Using Artificial Intelligence Techniques. 2010.
50.張善焜, 基於物件導向資料庫的屬性約簡系統研究. 2009, 朝陽科技大學資訊管理系.
51.蔡明富, 以分層抽樣之規則歸網探勘信用卡族群共同特性. 2005, 東海大學資訊工程與科學研究所.
52.黃代鈞, 以遺傳演算法結合貝氏分類法快速篩選與遺傳疾病相關的基因. 2003, 長庚大學資訊管理研究所碩士論文.
53.楊雅如, 類別屬性關聯性考驗離散化演算法. 2005, 國立台南大學資訊教育研究所.
54.Fayyad and Irani. Multi-Interval Discretization of Continuous-Valued Attributes for Classification Learning. 1993.
55.Peter D. Grünwald, I.J.M.a.M.A.P., Advances in Minimum Description Length Theory and Applications. 2005: The MIT Press.
56.林天和, 應用Fuzzy ARTMAP與Minimum Description Length Principle於臨床資料的學習與預測. 1996, 清華大學資訊工程學系.