隨著製造技術日益進步，以及資訊科技的蓬勃發展，面對瞬息萬變且競爭激烈的市場，企業須有效提升產品品質並降低成本，釐清影響製程之關鍵，以提升競爭力。然而，單憑工程知識或經驗法則已無法有效改善，因此本研究結合鑄造產業知識與資訊科技的應用，掌握關鍵的競爭優勢，進而提升企業產品品質。本研究應用資料探勘技術於鑄造業製程品質改善問題，提出一套屬性篩選流程，綜合類神經網路、隨機森林、支持向量迴歸、約略集合理論、迴歸分析等五種方法，彙整各方法篩選之重要變數，找出影響製程關鍵參數，並建構縮減模型。在確定關鍵製程參數後，結合類神經網路與基因演算法於製程參數最佳化，找出關鍵參數的最佳化參數組合。本研究以台灣某鑄造工廠為例，依照所提出之屬性篩選流程進行實證研究，將原先17項製程參數篩選出9項關鍵參數，並找出關鍵參數之最佳化參數組合。另外，比較五種資料探勘方法之縮減模型績效，結果顯示在屬性篩選後的縮減模型仍保有不錯的預測能力，說明本研究所提出屬性篩選流程之可行性。

With the advancement of manufacturing technology and the flourishing development of information technology, casting industry is faced with the increasingly competitive market, so companies must enhance their product’s quality and reduce manufacturing costs, and clarify what highly influences the process to have the key competitive advantage. However, simply relying on domain knowledge or rules of thumb is unable to identify the root causes of quality problems effectively.
This study applies data mining techniques for the process improvement issue of casting industry and proposes a general procedure for attribute selection. Five data mining techniques, including artificial neural network (ANN), random forest (RF), support vector machine (SVM), rough set theory (RST), and regression analysis are used to select the important attributes. This study aggregates the results from each method to identify the key parameters and builds the reduced model. In the end, the artificial neural network and genetic algorithm (GA) are utilized for optimizing the selected process parameters.The proposed procedure was employed to analyze the manufacturing data of a casting company in Taiwan. The research results presented that nine key process parameters were identified from seventeen original attributes and then the optimal combination of key parameters was obtained. In addition, the reduced model still maintained the exceptional ability to perform adequately, which confirmed the feasibility of the proposed attribute screening procedure.

摘要 I
Abstract II
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 3
第二章 文獻探討 4
2.1 鑄造產業 4
2.1.1 鑄造技術 4
2.1.2 鑄造之產業結構與發展 5
2.2資料探勘 6
2.2.1 資料探勘技術 6
2.2.2 資料探勘功能 8
2.2.3 資料探勘在鑄造產業之應用 9
2.3 屬性篩選過程與方法 9
2.3.1 類神經網路 10
2.3.2 隨機森林 11
2.3.3 支持向量迴歸 13
2.3.4 約略集合理論 14
2.3.5 迴歸分析 15
2.3.6 屬性篩選之應用 16
2.4 基因演算法 19
2.5 結合類神經演算法和基因演算法於參數設計之應用 21
第三章 研究方法 23
3.1 研究流程 23
3.2 資料前處理 25
3.3 屬性篩選之各方法模型建構 26
3.3.1 類神經網路模型建構 26
3.3.2 隨機森林模型建構 29
3.3.3 支持向量迴歸模型建構 30
3.3.4 約略集合理論模型建構 32
3.3.5 迴歸分析模型建構 36
3.4 結合類神經演算法和基因演算法於參數設計之應用 37
第四章 研究結果 39
4.1 個案描述與問題定義 39
4.2 資料蒐集與觀察 40
4.3 資料前處理 40
4.4 定義屬性 41
4.5 屬性篩選執行 41
4.5.1 類神經網路之屬性篩選 42
4.5.2 隨機森林之屬性篩選 46
4.5.3 支持向量迴歸之屬性篩選 48
4.5.4 約略集合理論之屬性篩選 51
4.5.5 迴歸分析之屬性篩選 52
4.5.6 重要屬性彙總 53
4.6 結合類神經網路與基因演算法於製程參數最佳化 54
第五章 結論 58
5.1 結論 58
5.2 未來研究方向 59
參考文獻 60

[1]Berson, A., Smith, S., & Thearling, K. (1999). An overview of data mining techniques: excerpted from the book building data mining applications for CRM. McGraw-Hill, 89-229.
[2]Breiman, L. (2001a). Random forests. Machine Learning, 45(1), 5-32.
[3]Breiman, L. (2001b). Statistical modeling: The Two Cultures, Statistical Science, 16, 199 -215.
[4]Chen, K. Y., & Wang, C. H. (2007). Support vector regression with genetic algorithms in forecasting tourism demand. Tourism Management, 28(1), 215-226.
[5]Chien, C. F., & Chen, L. F. (2008). Data mining to improve personnel selection and enhance human capital: a case study in high-technology industry. Expert Systems With Applications, 34(1), 280-290.
[6]Fayyad, U., & Stolorz, P. (1997). Data mining and KDD: promise and challenges. Future Generation Computer Systems, 13(2-3), 99-115.
[7]Fayyad, U., Piatetsky-Shapiro, G., & Smyth, P. (1996a). From data mining to knowledge discovery in databases. AI Magazine, 17(3), 37.
[8]Fayyad, U., Piatetsky-Shapiro, G., & Smyth, P. (1996b). The KDD process for extracting useful knowledge from volumes of data. Communications of the ACM, 39(11), 27-34.
[9]Goldberg, D. E. (1989). Genetic algorithms in search, optimization, and machine learning. Reading, MA: Addison-Wesley.
[10]Górny, Z., Kluska-Nawarecka, S., Wilk-Kołodziejczyk, D., & Regulski, K. (2010). Diagnosis of casting defects using uncertain and incomplete knowledge. Archives of Metallurgy and Materials, 55(3), 827-836.
[11]Hall, C. (1995). The devil's in the details: techniques, tools, and application for database mining and knowledge discovery part II. Intelligent Software Strategies, 6(9), 1-16.
[12]Han, J., Pei, J., & Kamber, M. (2011). Data mining: concepts and techniques. Elsevier.
[13]He, W., Wang, Z., & Jiang, H. (2008). Model optimizing and feature selecting for support vector regression in time series forecasting. Neurocomputing, 72(1), 600-611.
[14]Holland, J. H. (1975). Adaptation in natural and artificial systems. An Introductory Analysis with Application to Biology, Control, and Artificial Intelligence. University of Michigan Press.
[15]Hui, S. C., & Jha, G. (2000). Data mining for customer service support. Information & Management, 38(1), 1-13.
[16]Jack, L. B., & Nandi, A. K. (2000). Genetic algorithms for feature selection in machine condition monitoring with vibration signals. IEE Proceedings-Vision, Image and Signal Processing, 147(3), 205-212.
[17]Karimi, H., & Yousefi, F. (2012). Application of artificial neural network–genetic algorithm (ANN–GA) to correlation of density in nanofluids. Fluid Phase Equilibria, 336, 79-83.
[18]Khan, M. S., Muyeba, M., & Coenen, F. (2008). A weighted utility framework for mining association rules. In Computer Modeling and Simulation. EMS'08. Second UKSIM European Symposium on IEEE, 87-92.
[19]Lakshmanan S. (2010). Improving quality of sand casting using Taguchi method and ANN analysis. International Journal on Design and Manufacturing Technologies, 4(1).
[20]Oreski, S., & Oreski, G. (2014). Genetic algorithm-based heuristic for feature selection in credit risk assessment. Expert Systems with Applications, 41(4), 2052-2064.
[21]Pal, M., & Mather, P. M. (2003). An assessment of the effectiveness of decision tree methods for land cover classification. Remote Sensing of Environment, 86(4), 554-565.
[22]Penny, J., & Lindfield, G. (1995). Numerical methods using MATLAB. New York Ellis Horwood.
[23] Pawlak, Z. (1982). Rough sets. International Journal of Parallel Programming, 11(5), 341-356.
[24]Qian, Y., Liang, J., Pedrycz, W., & Dang, C. (2010). Positive approximation: an accelerator for attribute reduction in rough set theory. Artificial Intelligence, 174(9-10), 597-618.
[25]Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1985). Learning internal representations by error propagation (No. ICS-8506). California Univ San Diego La Jolla Inst for Cognitive Science.
[26]Schölkopf, B., & Smola, A. J. (2002). Learning with kernels: support vector machines, regularization, optimization, and beyond. MIT press.
[27]Shen, C., Wang, L., & Li, Q. (2007). Optimization of injection molding process parameters using combination of artificial neural network and genetic algorithm method. Journal of Materials Processing Technology, 183(2), 412-418.
[28]Sousa, S. I. V., Martins, F. G., Alvim-Ferraz, M. C. M., & Pereira, M. C. (2007). Multiple linear regression and artificial neural networks based on principal components to predict ozone concentrations. Environmental Modelling & Software, 22(1), 97-103.
[29]Su, C. T., Hsu, H. H., & Tsai, C. H. (2002). Knowledge mining from trained neural networks. Journal of Computer Information Systems, 42(4), 61-70.
[30]Su, C. T., Wang, P. C., Chen, Y. C., & Chen, L. F. (2012). Data mining techniques for assisting the diagnosis of pressure ulcer development in surgical patients. Journal of Medical Systems, 36(4), 2387-2399.
[31]Swiniarski, R. W., & Skowron, A. (2003). Rough set methods in feature selection and recognition. Pattern Recognition Letters, 24(6), 833-849.
[32]Thakare, A. G. & Tidke, D. J. (2013). Data mining for casting defects analysis. International Journal of Engineering Research & Technology, 12(2).
[33]Tsai, C. F., & Hsiao, Y. C. (2010). Combining multiple feature selection methods for stock prediction: union, intersection, and multi-intersection approaches. Decision Support Systems, 50(1), 258-269.
[34]Vapnik, V. (1998). Statistical learning theory. New York: Wiley Inc.
[35]Vapnik, V. N. (1995). The nature of statistical learning theory. New York, USA: Springer-Verlag.
[36]Vapnik, V., & Chapelle, O. (1999). Bounds on error expectation for SVM. MIT Press.
[37]Vapnik, V., Golowich, S. E., & Smola, A. J. (1997). Support vector method for function approximation, regression estimation and signal processing. In Advances in Neural Information Processing Systems, 281-287.
[38]Walczak, B., & Massart, D. L. (1999). Rough sets theory. Chemometrics and Intelligent Laboratory Systems, 47(1), 1-16.
[39]Zheng, B., Yoon, S. W., & Lam, S. S. (2014). Breast cancer diagnosis based on feature extraction using a hybrid of K-means and support vector machine
algorithms. Expert Systems with Applications, 41(4), 1476-1482.
[40]Zhou, C. C., Yin, G. F., & Hu, X. B. (2009). Multi-objective optimization of
material selection for sustainable products: artificial neural networks and genetic
algorithm approach. Materials & Design, 30(4), 1209-1215.
[41]蘇木春、張孝德 (2004)。機器學習：類神經網路、模糊系統以及基因演算法則(修訂二版)。台灣：全華圖書。
[42]蘇朝墩 (2013)。品質工程：線外方法與應用。前程文化事業有限公司。