運用間隔時間序列探勘於史蒂芬強生症候群病患罹病前之用藥序列與住院長短之關聯分析

簡易檢索 / 詳目顯示

回結果列表

研究生：	張之倩 Chih-Chien Chang
論文名稱：	運用間隔時間序列探勘於史蒂芬強生症候群病患罹病前之用藥序列與住院長短之關聯分析 Applying Time Interval Sequential Pattern Mining Approach to Investigate the Patients’ Medication Sequence and Length of Stays Before Suffering Steven Johnson Syndrome
指導教授：	歐陽超 Chao Ou-Yang
口試委員:	歐陽超 Chao Ou-Yang 郭人介 Ren-Jieh Kuo 汪漢澄 Han-Cheng Wang
學位類別：	碩士 Master
系所名稱：	管理學院 - 工業管理系 Department of Industrial Management
論文出版年：	2017
畢業學年度：	105
語文別：	中文
論文頁數：	55
中文關鍵詞：	史蒂芬強生症候群、高頻間隔時間序列探勘、K-means 群集分析、重複用藥、藥物交互作用
外文關鍵詞：	Stevens Johnson Syndrome, High Frequency Time Interval Sequential Pattern Mining, K-means, Duplicate Medications, Drug Interaction
相關次數：	點閱：225 下載：1
分享至:	分享至facebook 分享至twitter

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

國內醫療資源豐富，患者就醫方便，容易造成患者一種疾病在多家醫院就診，或
同時患有多種疾病分別看不同科別造成重覆用藥或多重藥物交互作用。另國內藥害
救濟給付案，以史蒂芬強生症候群(Stevens Johnson Syndrome;以下簡稱為SJS)症狀
占了申請的藥害救濟件比率最高。因此本研究希望能運用間隔時間為基之序列演
算法探討史帝芬強生症候群病患罹病前之拿藥序列。
傳統的序列探勘僅能得知項目間的先後發生順序，卻無法得知項目間的間隔時
間，因此在探勘有關SJS 患者的序列時，就會忽略藥物藥性未完全消逝，卻又跟著下
次拿藥的藥性產生交互作用影響，可能導致SJS 患者發病機率增加，因此本研究將
考慮間隔時間，將傳統序列探勘演算法加入拿藥間隔時間，探討SJS 患者住院前拿
藥間隔時間序列。
本研究個案使用台灣全民健保資料庫，參考某一年份SJS 患者的就醫記錄，串
聯資料庫中相關資料並將資料進行前處理。方便日後序列進行作業，將藥理類別、入
院及出院時間依據醫令進行分類編碼。而SJS 患者於確診日前的拿藥間隔時間與住
院時間利用K-means 群集分析方法，分別將此兩者連續型天數時間各自類別化。接
著透過高頻間隔時間序列演算法，從資料庫中挖掘出頻繁的間隔時間序列，找出患者
拿藥間隔時間長短與住院天數間的序列性關聯。經由本研究提出參考資訊作為醫院
醫療決策的輔助資訊，達到及早預防成效結果，為更多潛在患者帶來新的福音。

With plenty of convenient medical resources, it is easy for patients to see a doctor
in several hospitals with the same disease or let patients get duplicated medication or
interactions between multiple drugs. On the other hand, in the investigation of Taiwan
Drug relief system, it indicated that Stevens Johnson Syndrome (shortly as SJS in the
following paper) was the main case for application. Therefore, the aim of this study is
to use the time interval sequential pattern to explore the patients’ medication sequence
and length of stays before suffering Steven Johnson Syndrome.
However, the conventional sequential pattern mining methods only provide the
chronological order but it will ignore the time-interval between the items. For this
reason, using the approach to mine the patients who got SJS, it might lead to neglecting
the interactions between the medicines which would cause an increase in morbidity of
SJS. Consequently, the research of this study will consider the influences of time
interval in the above methods to discuss the time-interval sequential pattern of patients
who suffered from SJS.
This research has retrieved patients’ medical records from Taiwan National Health
Insurance Research Database and connected the relative data for preprocessing. To
easily analyze the data, pharmacological classification, admission and discharge were
classified and coded. In addition, K-means approach was initially adopted to analyze
the time interval between taking medicines and length of stays to categorize in
consecutive days. After that, we can apply high frequency pattern from the database to
discover correlation between receiving medicines and length of stays with time interval
sequential pattern. The results of the research can be regards as a reference for medical
decision making and a benefit to the potential patients.

摘 要............................................................................................................................ I
Abstract ......................................................................................................................... II
誌 謝.......................................................................................................................... III
目 錄.......................................................................................................................... IV
圖目錄.......................................................................................................................... VI
表目錄........................................................................................................................ VII
第一章、緒論................................................................................................................ 1
1.1 研究背景.......................................................................................................... 1
1.2 研究目的.......................................................................................................... 2
1.3 本計畫之重要性.............................................................................................. 3
第二章、文獻探討........................................................................................................ 4
2.1 史蒂芬強生症候群.......................................................................................... 4
2.2 資料探勘.......................................................................................................... 6
2.3 K-means 群集分析方法 .................................................................................. 7
2.4 序列型樣及間隔時間序列型樣探勘.............................................................. 8
2.5 關聯規則演算法.............................................................................................. 9
2.6 關聯規則衡量指標........................................................................................ 10
第三章、研究方法...................................................................................................... 11
3.1 研究流程與架構............................................................................................ 11
3.2 資料前處理.................................................................................................... 13
3.2.1 健保資料庫類別檔整合..................................................................... 13
3.2.2 篩選SJS 患者資料............................................................................. 14
V
3.2.3 用藥分類及以醫令作分類編碼......................................................... 15
3.2.4 部份資料連續型屬性類別化............................................................ 18
3.3 序列型樣探勘................................................................................................ 19
3.4 高頻間隔時間序列演算法............................................................................ 24
3.5 高頻拿藥間隔時間序列關聯規則................................................................ 30
第四章、實作研究...................................................................................................... 31
4.1 資料介紹........................................................................................................ 31
4.2 資料前處理.................................................................................................... 32
4.2.1 健保資料庫類別檔整合實作............................................................. 32
4.2.2 篩選患者SJS 資料............................................................................. 34
4.2.3 患者用藥項目篩選及編碼................................................................. 35
4.2.4 部份資料連續型屬性類別化............................................................ 37
4.3 高頻間隔時間序列........................................................................................ 37
4.4 高頻間隔時間序列關聯規則........................................................................ 39
4.5 結果分析與探討............................................................................................ 41
第五章、結論與建議.................................................................................................. 43
5.1 結論................................................................................................................ 43
5.2 未來建議....................................................................................................... 44
參考文獻...................................................................................................................... 45
附錄(一) ....................................................................................................................... 48
附錄(二) ....................................................................................................................... 49
附錄(三) ....................................................................................................................... 52
                                

Agrawal, R., Imieliński, T., & Swami, A. (1993). Mining association rules between sets
of items in large databases. Paper presented at the Acm sigmod record.
Agrawal, R., & Srikant, R. (1995). Mining sequential patterns. Paper presented at the
Data Engineering, 1995. Proceedings of the Eleventh International Conference
on.
Chen, Y.-L., Chiang, M.-C., & Ko, M.-T. (2003). Discovering time-interval sequential
patterns in sequence databases. Expert Systems with Applications, 25(3), 343-
354.
Cheng, C.-Y. (2016). Applying Nested Genetic Algorithm to Discover the Patterns of
Association Rules of Cerebrovascular Health Examinees' Re-coming.
Chu, H.-Y. (2016). Applying Sequential Pattern Approach to Identify Clinic Sequence-
An Example of Stevens-Johnson Syndrome.
Chung, W.-H., Hung, S.-I., Hong, H.-S., Hsih, M.-S., Yang, L.-C., Ho, H.-C., . . . Chen,
Y.-T. (2004). Medical genetics: a marker for Stevens–Johnson syndrome.
Nature, 428(6982), 486-486.
Fayyd, U. M., Shapiro, G. P., & Smyth, P. (1996). From data mining to knowledge
discovery: an overview.
Frawley, W. J., Piatetsky-Shapiro, G., & Matheus, C. J. (1992). Knowledge discovery
in databases: An overview. AI magazine, 13(3), 57.
Grupe, F. H., & Mehdi Owrang, M. (1995). Data base mining discovering new
knowledge and competitive advantage. Information System Management,
12(4), 26-31.
Kanungo, T., Mount, D. M., Netanyahu, N. S., Piatko, C. D., Silverman, R., & Wu, A.
Y. (2002). An efficient k-means clustering algorithm: Analysis and
implementation. IEEE transactions on pattern analysis and machine
intelligence, 24(7), 881-892.
Karabatak, M., & Ince, M. C. (2009). An expert system for detection of breast cancer
based on association rules and neural network. Expert Systems with
Applications, 36(2), 3465-3469.
Leenutaphong V, Sivayathorn A, Suthipinittharm P, Sunthonpalin P. (1993). Stevens—
Johnson syndrome and toxic epidermal necrolysis in Thailand. International
journal of dermatology; 32: 428–31.
Linoff, G. S., & Berry, M. J. (2011). Data mining techniques: for marketing, sales, and
customer relationship management: John Wiley & Sons.
MacQueen, J. (1967). 5th Berkeley symposium on mathematical statistics and
46
probability. Berkeley, CA.
Ordonez, C., Omiecinski, E., De Braal, L., Santana, C. A., Ezquerra, N., Taboada, J.
A., . . . Garcia, E. V. (2001). Mining constrained association rules to predict
heart disease. Paper presented at the Data Mining, 2001. ICDM 2001,
Proceedings IEEE International Conference on.
Rallison, M., Carlisle, J., Lee, R., Vernier, R., & Good, R. (1961). Lupus erythematosus
and Stevens-Johnson syndrome: occurrence as reactions to anticonvulsant
medication. American Journal of Diseases of Children, 101(6), 725-738.
Rzany, B., Correia, O., Kelly, J. P., Naldi, L., Auquier, A., & Stern, R. (1999). Risk of
Stevens-Johnson syndrome and toxic epider mal necrolysis during first weeks
of antiepileptic therapy: a case-control study. The lancet, 353(9171), 2190-2194.
Strom, B. L., Carson, J. L., Halpern, A. C., Schinnar, R., Snyder, E. S., Shaw, M., . . .
Chen, D. (1991). A population-based study of Stevens-Johnson syndrome:
incidence and antecedent drug exposures. Archives of dermatology, 127(6),
831-838.
Tennis, P., & Stern, R. S. (1997). Risk of serious cutaneous disorders after initiation of
use of phenytoin, carbamazepine, or sodium valproate: a record linkage study.
Neurology, 49(2), 542-546.
Tigchelaar, H., Kannikeswaran, N., & Kamat, D. (2008). Stevens-Johnson syndrome:
An intriguing diagnosis. Consultant for Pediatricians.
Trivedi, B. S., Darji, N. H., Malhotra, S. D., & Patel, P. R. (2016). Antiepileptic drugsinduced
Stevens–Johnson syndrome: A case Series. Journal of basic and
clinical pharmacy, 8(1), 42.
Tsai, C.-Y., & Chen, I.-C. (2009). Using Genetic Process Mining Technology to
Construct a Time-Interval Process Model. Paper presented at the International
Conference on Industrial, Engineering and Other Applications of Applied
Intelligent Systems.
Ward, K. E., Archambault, R., & Mersfelder, T. L. (2010). Severe adverse skin reactions
to nonsteroidal antiinflammatory drugs: A review of the literature. American
Journal of Health-System Pharmacy, 67(3).
Wu, S.-Y., & Chen, Y.-L. (2007). Mining nonambiguous temporal patterns for intervalbased
events. IEEE transactions on knowledge and data engineering, 19(6).
Yang, C.-Y. (2014). A Hybrid of Data Mining and Statistical Analysis Approach on
Association between Pulmonary Tuberculosis and Lung Cancer.
Yen, S.-J., Lee, Y.-S., Shie, B.-E., & Lee, Y.-K. (2015). Mining Sequential Patterns
with Pattern Constraint. Paper presented at the Asian Conference on Intelligent
Information and Database Systems.
47
台灣財團法人藥害救濟基金會。取自:http://www.tdrf.org.tw/
江美靜（2002）。有時間區間的循序挖掘。國立中央大學資訊管理研究所。
姚志成（2005）。運用資料探勘技術建構脂肪肝預測模式。中原大學資訊管理研
究所。
若天妤（2015）。運用關聯分析法探討腦部健診民眾回診型態。國立臺灣科技大
學工業管理系。
陳品文（2007）。K-means 群集演算法初始化之新方法。國立東華大學企業管理
學系。
翁靜美（2011）。在資料串流的環境下探勘高效益項目集之研究。銘傳大學資訊
工程學系碩士班。
張志揚（2003）。挖掘不同時間單位之時間序列型樣。國立臺灣科技大學資訊管
理系。
梁忻琳（2009）。台灣嚴重藥物不良反應-史蒂文生氏強生症候群致死案例與藥物
交互作用相關性研究。高雄醫學大學藥學研究所碩士在職專班。
張仁鴻, 郭樂明, 賴豐傑, 陳勝咸, 孫啟書, 莊世松, . . . 王建楠. (2010).
Mefenamic Acid 導致罕見的肝細胞混合膽汁鬱積性肝炎及史帝文生強生
症候群 : 一病例報告. 內科學誌, 21 卷(1 期), 頁66-71.
游雅雯, 鄭博文, 林宏茂, & 梁玉芬. (2012). 以資料探勘技術預測健康檢查大腸
息肉之風險因子. 醫務管理期刊, 13 卷(3 期), 頁162-178.
曾冠蓉（2016）。重複用藥傷身體，雲端藥歷守護您.恩主公醫訊, Nov.2016 , 頁
5-7
楊倍權（2013）。從不確定資料庫探勘頻繁項目集。銘傳大學資訊工程學系碩士
班。
鄭昕弦, 張豫立, & 周月卿. (2013). 疑似Allopurinol 引起史蒂文生強生症候群.
醫院藥學, 30 卷(1/2), 頁43-48.
鄭智元（2016）。運用巢狀基因演算法於腦部健檢民眾回診關聯規則模式之探討。
國立臺灣科技大學工業管理系。
賴昆汕（2004）。藥品交互作用資料庫應用系統設計與實作。國立臺灣大學資訊
工程學研究所。
謝百恩（2006）。有型樣限制的序列型樣探勘演算法。銘傳大學資訊工程學系碩
士班。
簡君如（2010）。應用序列樣式探勘技術建構一個時間間隔序列最佳化分類模型。
元智大學工業工程與管理學系。
藍國誠 (2009)。運用資料探勘技術建構健康趨勢及疾病關聯性之分析預測系統。
國立成功大學資訊工程研究所

簡易檢索 / 詳目顯示

相關論文