院內心跳停止(In Hospital Cardiac Arrest, IHCA)一直是醫院的重要議題，而國內發生IHCA後實施急救措施後的出院率和存活率遠低於美國。急救措施的執行成效關係到能否立即拯救危急瀕死病患之生命，過去少有應用資料挖礦方法論於IHCA之研究，使用新穎的資料挖掘方法論於IHCA案例上是值得研究的議題。
本研究提出整合之方法，應用SVM產生的SV結合規則萃取演算法如決策樹C5.0、CART以及粗糙集方法論，以增加SVM的解釋能力與萃取出感興趣的資訊，並且具備良好的分類性能。
本研究運用五個指標，準確度(Accuracy)、敏感度(Sensitivity)、特異度(Specificity)、保真度(Fidelity)和覆蓋率(Coverage)指標來評估所提出的三種整合型方法在10個UCI的資料集和醫療案例上的分類性能和解釋能力。研究結果顯示本研究所討論的三種方法皆可以萃取出80%以上的整體準確度之規則，說明本研究可以成功運用在UCI以及IHCA實例上，使得發生IHCA患者的出院率大幅提升以及醫療資源能夠更為妥當的利用。

In-Hospital Cardiac Arrest (IHCA) has been an important issue in hospitals, and the IHCA patients in Taiwan have much lower survival and discharge rate than in the United States. The effectiveness of resuscitation may ultimately determine whether a patient can survive or discharge from the hospital. The data mining methods have rarely been applied to IHCA cases and to extract useful rules to assist medical personnel in decision making. Therefore, it is desired to study the application of support vector machines (SVM) in integrating a set of methods to provide a better understanding of the relationship between IHCA and discharge rate for the hospital personnel.
To solve the problem of insufficient explanatory power of SVM, this study integrates three rule extraction algorithms, the C5.0 decision trees, classification and regression trees (CART), and rough sets, with the support vectors generated from SVM to extract rules and enhance SVM's explanatory power. Five indexes including accuracy, sensitivity, specificity, fidelity, and coverage are used to evaluate the classification performance of the three proposed methods on the UCI database and IHCA case.
The results show that all three proposed methods can obtain more than 80% overall accuracy of classification. Thus, the proposed methods not only can be applied successfully on the UCI dataset, but also can provide useful reference for medical decisions in the real case of IHCA.

摘要 i
目錄 iii
第一章 緒論 1
1.1 研究背景與動機 1
1.2 研究目的 2
1.3 研究架構 2
第二章 文獻探討 5
2.1 院內心跳停止 5
2.2 快速反應小組(rapid response team, RRT) 6
2.3 預警系統(early warning system, EWS) 9
2.4 支持向量機 10
2.5 決策樹 11
2.5.1 ID3, C4.5 & C5.0 13
2.5.2 CART (Classification and Regression Tree) 14
2.6 粗糙集方法論 (Rough sets) 15
2.7 規則萃取 17
2.7.1 SVM + Prototypes 18
2.7.2 RulExSVM 19
第三章 方法論 21
3.1 研究方法概念 21
3.2 SVM + C5.0 25
3.2 SVM + CART 25
3.2 SVM + Rough sets 26
第四章 數據分析 28
4.1 研究平台及資料處理步驟 28
4.2 分類指標 28
4.3 數據 30
4.4 分析結果 31
第五章 案例研究 37
5.1 研究背景 37
5.2 案例分析 38
5.2.1 SVM + C5.0 38
5.2.2 SVM + CART 41
5.2.3 SVM + Rough sets 42
5.2.4總結 44
第六章 結論 45
6.1 結論 45
6.2 未來展望 46
參考文獻 47

林昕潔(2005年)。以SVM與詮釋資料設計書籍分類系統。國立交通大學資訊科學與工程碩士論文。
財團法人醫院評鑑暨醫院品質策進會(2013年)。2013醫療品質突破系列-拯救生命大作戰:提升院內心跳停止照護品質說明。臺北：醫策會。
陳衍成(2012年)。支持向量器的分類和規則萃取：理論與運用。國立清華大學工業工程與工程管理學系博士論文。
溫坤禮(2008年）。粗糙集入門及應用。台北市:五南書出版股份有限公司。
謝邦昌(2005年)。資料採礦與商業智慧-SQL Server 2005。台北市：鼎茂圖書出版股份有限公司。
American Heart Association Statistics Committee and Stroke Statistics Subcommittee (2013).Heart disease and stroke statistics--2013 update: a report from the American Heart Association. Circulation, 113(6). doi: 10.1161/CIR.0b013e31828124ad
Barakat N., Bradley A. P. (2007). Rule Extraction from Support Vector Machines: A Sequential Covering Approach. IEEE Transactions on Knowledge and Data Engineering, 19(6), 729-741.
Bellomo, R., Goldsmith D., Uchino S., Buckmaster J., Hart G., Opdam H., . . . Gutteridge G. (2004). Prospective controlled trial of effect of medical emergency team on postoperative morbidity and mortality rates. Critical Care Medicine, 32(4), 916-921. doi:10.1097/01.ccm.0000119428.02968.9e.
Breiman L., Friedman J., Charles J., StoneR.A. Olshen (1984). Classification and regression trees (1st ed.). CA: Chapman and Hall/CRC.
Buist, M. D. (2002). Effects of a medical emergency team on reduction of incidence of and mortality from unexpected cardiac arrests in hospital: preliminary study. Bmj, 324(7334), 387-390. doi:10.1136/bmj.324.7334.387.
Burch V. C., Tarr G., & Morroni C. (2008). Modified early warning score predicts the need for hospital admission and inhospital mortality. Emergency Medicine Journal, 25(10), 674-678. doi:10.1136/emj.2007.057661.
Cormen T. H., Leiserson C. E., & Rivest R. L. (1990). Introduction to algorithms. Cambridge, MA: MIT Press.
Cortes C., Vapnik V. (1995). Support-vector networks. Machine Learning, 20 (3), 273–297. doi:10.1007/BF00994018.
Curram S., Mingers J. (1994). Neural Networks, Decision Tree Induction and Discriminant Analysis: An Empirical Comparison. The Journal of the Operational Research Society, 45(4), 440-450. doi:10.2307/2584215
Debard M. L. (1981). Cardiopulmonary resuscitation: Analysis of six years' experience and review of the literature. Annals of Emergency Medicine, 10(8), 408-416. doi:10.1016/s0196-0644(81)80307-1.
Dosh K., Dhoble A., Evonich R., Gupta A., Shah I., Gardiner J., & Dwamena F. C. (2009). Analysis of limited resuscitations in patients suffering in-hospital cardiac arrest. Resuscitation, 80(9), 985-989. doi:10.1016/j.resuscitation.2009.05.011.
Esposito F., Malerba D., Semeraro G., & Kay J. (1997). A comparative analysis of methods for pruning decision trees. IEEE Transactions on Pattern Analysis and Machine Intelligence, 19(5), 476-493. doi:10.1109/34.589207.
Forestier G., Gançarski P., & Wemmert C. (2010). Collaborative clustering with background knowledge. Data & Knowledge Engineering, 69(2), 211-228. doi:10.1016/j.datak.2009.10.004.
Fu X., Ong C., Keerthi S., Hung G. G., & Goh L. (2004). Extracting the knowledge embedded in support vector machines. 2004 IEEE International Joint Conference on Neural Networks (IEEE Cat. No.04CH37541). doi:10.1109/ijcnn.2004.1379916
Gretton A., Doucet A., Herbrich R., Rayner P., & Scholkopf B. (2001). Support vector regression for black-box system identification. Proceedings of the 11th IEEE Signal Processing Workshop on Statistical Signal Processing (Cat. No.01TH8563). doi:10.1109/ssp.2001.955292.
Grzymala-Busse J. W. (2005). Rough Set Strategies to Data with Missing Attribute Values. Foundations and Novel Approaches in Data Mining Studies in Computational Intelligence, 197-212. doi:10.1007/11539827_11.
Hart G. K. (2001). Antecedents to hospital deaths: all in good time. Internal Medicine Journal, 31(6), 321-321. doi:10.1046/j.1445-5994.2001.00094.
Kouwenhoven W.B., Jude J.R., Knickerbocker G.G.(1960). Closed-chest cardiac massage. JAMA, 173(10):1064-1067. doi:10.1001/jama.1960.03020280004002
Hall L., Chawla N., & Bowyer K. (1998). Decision tree learning on very large data sets. SMC'98 Conference Proceedings. 1998 IEEE International Conference on Systems, Man, and Cybernetics (Cat. No.98CH36218). doi:10.1109/icsmc.1998.725047.
Joffe A.R., Anton N.R., Burkholder S.C.(2001). Rapid Response Team Implementation in a Children’s Hospital—Reply. Arch Pediatr Adolesc. Med,165(12):1139-1140. doi:10.1001/archpedi.165.12.1139-b
Leung F., Lam H., Ling S., & Tam P. (2003). Tuning of the structure and parameters of a neural network using an improved genetic algorithm. IEEE Transactions on Neural Networks, 14(1), 79-88. doi:10.1109/tnn.2002.804317.
Lian, H., & Lu, B. (2007). Multi-View Gender Classification Using Multi-Resolution Local Binary Patterns And Support Vector Machines. International Journal of Neural Systems, 17(06), 479-487. doi:10.1142/s0129065707001317
Loh W. (2014). Fifty Years of Classification and Regression Trees. International Statistical Review, 82(3), 329-348. doi:10.1111/insr.12016.
Mcgaughey J., Alderdice F., Fowler R., Kapila A., Mayhew A., & Moutray M. (2007). Outreach and Early Warning Systems (EWS) for the prevention of Intensive Care admission and death of critically ill adult patients on general hospital wards. Cochrane Database of Systematic Reviews. doi:10.1002/14651858.cd005529.pub2.
Mcgrath R. B. (1987). In-house cardiopulmonary resuscitation ? after a quarter of a century. Annals of Emergency Medicine, 16(12), 1365-1368. doi:10.1016/s0196-0644(87)80420-1.
Paterson R., Macleod D., Thetford D., Beattie A., Graham C., Lam S., & Bell D. (2006). Prediction of in-hospital mortality and length of stay using an early warning scoring system: clinical audit. Clinical Medicine, 6(3), 281-284. doi:10.7861/clinmedicine.6-3-281.
Pawlak Z. (1991). Rough sets: theoretical aspects of reasoning about data. Norwell, MA, USA:Kluwer Academic.
Nahla Barakat, Andrew P. Bradley(2009). Rule extraction from support vector machines: A review. Neurocomputing, 74(1-3), 178-190.
Thomas K., VanOyen Force M., Rasmussen D., Dodd D., Whildin S. (2007). Rapid response team: challenges, solutions, benefits.Crit Care Nurse,27(1):20-7.
Tseng T. & Huang C. (2007). Rough set-based approach to feature selection in customer relationship management. Omega, 35(4), 365-383. doi:10.1016/j.omega.2005.07.006.
Velayutham C., & Thangavel K. (2011). Unsupervised Feature Selection in digital mammogram image using rough set based entropy measure. 2011 World Congress on Information and Communication Technologies. doi:10.1109/wict.2011.6141318.
Walczak B., & Massart D. (1999). Rough sets theory. Chemometrics and Intelligent Laboratory Systems, 47(1), 1-16. doi:10.1016/s0169-7439(98)00200-7.
Wu X., & Kumar V. (2009). The top ten algorithms in data mining. Boca Raton: CRC Press.