玻璃鍍膜製造產業在現今競爭激烈的環境下，公司內部能否有效率地進行成本降低，已經是公司提升競爭力重要的一項決策。在產品製造過程中機台設備的屬性資料往往大量紀錄而無法有效利用，生產人員常藉著經驗的法則，進行生產，造成時間成本增加。本研究試圖從生產設備上紀錄的大量屬性資料，利用資料探勘的數值預測方法，建構出預測模型，來預測生產產品特性值中的反射率，玻璃顏色座標a與b值等重要的產品規格特性值，並找出那些屬性是影響產品特性值的重要屬性，幫助機台設備保養後或切換產品品種後進行生產時，能利用此模型，有效的預測產品特性與找出機台設備的重要屬性和調校關係，降低時間與人力成本。
本研究使用資料探勘的工具模式樹，分析產品生產中的鍍膜機設備大量資料，研究中各別對三個光學品質特性建立預測模型，研究結果顯示，反射率這組預測的誤差錯誤最低，也從中分析了這三個模式樹與其中的迴歸式，並找出影響抗反射薄膜產品的三項光學品質特性間的重要屬性。

Modern manufacturing industry of glass coating manufacturing industry is highly competitive, and the decision for efficiently reducing manufacturing costs is important for a company to enhance its competitiveness. The usage of the large amount of data collected from manufacturing processes is not effective, and hence the high production cost is generally caused by the decisions based on personal experience. This study attempts to use the data with plenty of attributes collected from production equipment to build model trees for predicting the characteristics of antireflection optical thin films, including reflection rate and the a and b values of glass color coordinates. The model trees can provide the critical attributes that will affect the characteristics of products. This kind of information can help engineers to set proper manufacturing parameters for reducing setup time and labor cost. The experimental results show that the prediction error of the reflectivity rate is the smallest, and the model tree for each of the three optical characteristics does suggest critical attributes for improving product quality.

中文
吳咏璁.(2004).模式樹選取分割屬性方法之研究. 成功大學工業與資訊管理學
系學位論文。
周君桂.(2016).使用模式樹於太陽能電池效率預測.成功大學工業與資訊管理學
系碩士在職專班學位論文。
李正中.(2009).薄膜光學與鍍膜技術(第六版).台北:華義圖書出版.
黃誌成.(2017).應用模式樹與偏最小平方法預測氧化因錫薄膜電阻值之研究.
成功大學工業與資訊管理學系碩士在職專班學位論文.
英文
Anderson, D. R., Sweeney, D. J., Williams, T.A., Chen, J.C. (2006). Statistics for Business and Economics: A practical approach, Thomson learning, Singapore.
Bogatin, E. (2004). Signal integrity - simplified. Upper saddle river, New Jersey: Pearson education, Inc..Figure 8-2 and Eqn. 8-1 Pg. 279.
Breiman, L., Friedman, J., Stone, C.J., & Olshen, R.A. (1984). Classification and Regression Tree. New York：Chapman and Hall.
Chen, S.-W., Chen, Y.-j., Wang, S. T., & Hwang, R.-C(2015). The Analysis of Influence Factor of ITO Bar’s Ideal Resistance for Multi-Finger Touch Screen. source：International journal of innovation science, Engineering & Technology, 2, 271-275.
Clarke, R. H. (1983). Theory of reflection from antireflection coatings. The Bell System Technical Journal. Vol.62, No. 10, December Printedin U.S.A.
Guild, J. (1931). The colorimetric properties of the spectrum. Philosophical transactions of the royal society of London., A230: 149–187.
Kutner, M. H., Nachtsheim, C., & Neter, J. (2004). Applied Linear Regression Models: McGraw-Hill/Irwin.
Lobur, M. Stekh, Yu. Kernytskyy, A. Faisal M. E. Sardieh. (2008). Some trends in knowledge discovery and data mining. International conference on perspective technologies and methods in MEMS design.
Manickavel, A S. A. Am1 Shalom V. Ravi* (2004). Sputter process variables prediction via data mining. IEEE Conference on cybernetics and intelligent systems Singapore.
Quinlan, J. R. (1992). Learning with continuous classes. Paper presented at the 5th Australian joint conference on artificial intelligence.
Wang, Y., & Witten, I. H. (1996). Induction of Model Tree for Predicting Continuous classes. Presented at the proceedings of the poster papers of the 9th European Conference on Machine Learning.
William J. Frawley, Gregory Piatetsky-Shapiro, and Christopher J. Matheus (1992). Knowledge discovery in databases: An overview .AI magazine volume 13 Number 3(© AAAI).