以決策樹法歸納關鍵製程暨以群集法識別關鍵路徑

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：20

、訪客IP：18.191.203.35

姓名

張筱芬(Hsiao-Fen Chang) 查詢紙本館藏

畢業系所

工業管理研究所在職專班

論文名稱

以決策樹法歸納關鍵製程暨以群集法識別關鍵路徑
(Critical Processes Induction with CART And Critical Path Identification with Clustering)

相關論文

★ 二階段作業研究模式於立體化設施規劃應用之探討–以半導體製造廠X及Y公司為例	★ 推行TPM活動以改善設備總合效率並提昇企業競爭力...以U公司桃園工廠為例
★ 資訊系統整合業者行銷通路策略之研究	★ 關鍵績效指標(KPI)之建立與推行 - 在造紙業
★ 應用實驗計劃法- 提昇IC載板錫球斷面品質最佳化之研究	★ 如何從歷史鑽孔Cp值導出新設計規則進而達到兼顧品質與降低生產成本目標
★ 產品資料管理系統建立及導入-以半導體IC封裝廠C公司為例	★ 企業由設計代工轉型為自有品牌之營運管理
★ 運用六標準差步驟與FMEA於塑膠射出成型之冷料改善研究(以S公司為例)	★ 台灣地區輪胎產業經營績效之研究
★ 以方法時間衡量法訂定OLED面板蒸鍍有機材料更換作業之時間標準	★ 利用六標準差管理提升生產效率－以Ａ公司塗料充填流程改善為例
★ 依流程相似度對目標群組做群集分析- 以航空發動機維修廠之自修工件為例	★ 設計鏈績效衡量指標建立 —以電動巴士產業A公司為例
★ 應用資料探勘尋找影響太陽能模組製程良率之因子研究	★ 整合性管理系統之建置研究 – 以A公司之管理系統整合為例

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 ( 永不開放)

摘要(中)

許多的品質工具以及手法被發展以及運用於各產業。這些工具以及手法協助組織確認出關鍵品質特性或是關鍵製程。這些方法也被廣泛運用於液晶顯示器產業。其中最為普遍運用的品質工具之一為失效模式分析，這項工具幫助組織確認出關鍵品質特性或是關鍵製程。此方法為藉由風險優先數的計算定義出關鍵品質特性或是關鍵製程，若此風險優先數為最大或是大於某特定數, 該項特性則視為關鍵品質特性或是關鍵製程。雖然對於RPN 之組成要素嚴重度(S)、發生度(O)、和難檢度(D) 有評分準則, 但此方法中評分結果仍被視為主觀的。
本研究主要為提議以資料探勘之方法論並自一本國面板廠商搜集相關資料。企圖以客戶聲音為目標, 將客戶聲音轉換成關鍵品質。我們運用了其中一項決策樹之方法, 分類及迴歸樹規納出影響每個批次產品品位之製程。進一步地，將導出之關鍵製程與每批次品位數以群集法分群並定一指標，藉由以此方法，可發現每一製程表現較好與較差的機台，關鍵製程中表現較好的機台便構成了關鍵路徑。

摘要(英)

Many quality tools and techniques have been developed and deployed extensively in the industry-wide to help the organization to identify and determine the significant characteristic of product and critical process in the organization. Those quality tools are adopted in the TFT-LCD industry too. One of the most commonly used techniques is the failure mode and effect analysis (FMEA) that helps to identify and determine the significant characteristic of products and critical process in the organizations. The key characteristic and critical process is determined if its RPN is the biggest or bigger than a specific one. Although there are guidelines for the scoring, the outcome is relatively subjective. Although there are guidelines for the scoring, the outcome is relatively subjectively.
This research is proposing another data mining procedure and collecting the dataset from a local LCD manufacturer. We attempt to aim at the Voice Of Customer (VOC), i.e. what customer requirements to our product are and transfer the VOC to critical to quality. The one of the decision tree techniques, Classification and Regression tree (CART) will be adopted. . With CART modeling, the decision tree inducts the processes that affect the total quantity of product rank per lot. Further, product rank are grouped and analyzed by the clustering model. Also we define an index to decide the equipment performance. Based on the result, the superior and the inferior equipments of each process are identified. The critical path is composed of the superior equipments. Meanwhile, the result can help the management to analyze the variation between equipments.

關鍵字(中)

★ 資料探勘
★ 決策樹
★ 群集法
★ 液晶顯示器

關鍵字(英)

★ data mining
★ decision tree
★ CART
★ Clustering
★ LCD

論文目次

Chinese Abstract I
English Abstract II
Acknowledgement III
Content IV
The Content of Figure VI
The Content of Table VII
1. Introduction 1
1.1 Background 1
1.2 Motivation 2
1.3 LCD Process 4
1.3.1. Array Process 4
1.3.2. CELL Process 6
1.3.3. Module Process 7
2. Literature Review 8
2.1 Failure Mode and Effect Analysis ( FMEA) 8
2.1.1. The objective of failure mode and effect analysis (FMEA) 8
2.1.2. Product, Process and Service characteristics 9
2.1.3. The Process of conducting an FMEA 11
2.1.4. The Challenge of RPN 13
2.2 Data Mining Technique 15
2.2.1. Data Mining introduction in brief 15
2.2.2. Basic Data mining tasks 16
3. Methodology 18
3.1 Classification and Regression Tree 18
3.1.1. Overview of CART 18
3.1.2. Splitting rule 19
3.1.3. Pruning 21
3.1.4. Researches of decision tree application in the manufacturing industry 22
3.2 K-means Clustering 23
3.2.1. Overview of K-means 23
3.2.2. K-means Algorithm 24
3.2.3. Researches with regard to Clustering 25
4. Critical Process induced by Data mining 26
4.1 Business understanding 26
4.2 Data Understanding 26
4.3 Data Preparation 27
4.4 Modeling and Evaluation 29
4.4.1. Critical Process determined by CART 29
4.4.2. Identify the Critical Production Path by Clustering 35
5. Summary 39
5.1 The Research Summary 39
5.2 Future Research 39
Reference 40

參考文獻

Reference
1. Breiman, L., J. H. Friedman, R. A. Olshen, and C.J. Stone, “Classification and Regression Trees” , : Wadsworth International Group, Belmont, Calif., U.S.A.,1984
2. Boris Mirkin ,“Clustering for data mining : a data recovery approach”, Chapman & Hall/CRC, Boca Raton, FL., U.S.A.,2005
3. Chia-Ming,Chang ,”Using CART Algorithm to develop the relation model between bin yield and WAT parameters “, Master Thesis, Graduate Industrial Engineering and Engineering Management, National Tsing Hua University, Republic Of China,2006
4. Chidanand Apté and Sholom Weiss, “ Data mining with decision trees and decision rules”, Future Generation Computer System Vol. 13,pp197-210.,1993
5. Chin-Sheng, Leu ,” Integrated Circuit Yield Estimation - Using Modified Poisson Model” Master Thesis, Department of Industrial Engineering and Management
National Chiao Tung University, Republic of China, 1993
6. Kun-Lin Hsieh,” The application of clustering analysis for the critical areas
on TFT-LCD panel”, Expert Systems with Applications Vol.34, pp952-957,2008
7. Hailiang Huang and Dianliang Wu “Product quality improvement analysis using data mining: A case study in ultra-precision manufacturing industry” Springer-Verlag LNAI 3614, pp. 577 – 580, Berlin Heidelberg, 2005
8. Hsin-Hung Kuo,” The application of data mining technology in LCD flat panel diagnosis” , Master Thesis, ,National Central University, Taiwan, 2007
9. Hsu-Cheng Fu,” Integrated Circuit Yield Model with Defect Source and Defect Clustering” , Master Thesis, Department of Industrial Engineering and Management
National Chiao Tung University, Republic of China, 1998
10. John B. Bowles,” An assessment of RPN prioritization in a failure modes effects and criticality analysis” Annual Reliability and maintainability Symposium.2003
11. Margaret H. Dunham,” Data mining introductory and advanced topics” Upper Saddle River, N.J. : Prentice Hall/Pearson Education,2003
12. M. Ben-Daya and Abdul Raouf,”A revised failure mode and effects analysis model” International Journal of Quality & Reliability Management, Vol 13 No.1 pp43-47,1996
13. Perry Lee , “ Constructing a semiconductor manufacturing data mining framework, developing a decision tree algorithm for classification and conducting empirical studies “
Master thesis, Graduate Industrial Engineering and Engineering Management Department, National Tsing Hua University, Taiwan, 2002
14. Robert Groth,” Data mining: building competitive advantage”, : Prentice Hall PTR ,Upper Saddle River, NJ. U.S.A.,1999
15. Stamatis, D. H.(1995), “ Failure mode and effect analysis : FMEA from theory to execution”,ASQC Quality Press ,Milwaukee, Wisc, U.S.A.
16. SPSS Inc.,“ Clementine10.1 Algorithms Guide “, ,Integral Solution Limited.
17. SPSS Inc.,“ Clementine10.1 Modeling Node “, ,Integral Solution Limited.
18. Maurizio Bevilacq, et al. “ The classification and regression tree approach to pump failure rate analysis” Reliability Engineering and System Safety Vol.79, pp50-6

指導教授

曾富祥(Fu-Shiang Tseng)

審核日期

2009-5-23

推文