全球半導體製造工藝日益精進，晶圓製程愈來愈微小，對於晶圓檢測設備是一大挑戰，而電子束檢測設備近年來備受關注，它能檢測微小晶圓，並透過圖像堆疊顯示缺陷，但其操作難度高導致產品失效狀況仍層出不窮，如何找出關鍵因子，降低失效問題，給予客戶高品質產品便是當前最緊迫的難題。本研究以南部科學園區某公司的電子束檢測設備為案例，於2015年1月至2021年6月，為期78個月共蒐集503筆出貨量，將偏轉圖像控制的母體資料進行資料探勘後，先透過SPSS軟體進行主成分分析，萃取出最具相關性與共同性的類別變數，再進行邏輯回歸、決策樹以及類神經網路三種機器學習，找出二元分類衡量指標、ROC（Receiver operating characteristic curve）曲線下面積AUC (Area under the Curve）及機器學習準確率高之4大項測試項目分類變數中，皆有重複出現的類別變數，交叉驗證後，得到兩項關鍵類別變數為電流放大增益倍數Gain與灰階值線性度偏移Offset，結果發現兩項類別變數和原先研發部門所設定的6大項測試分類變數做結合進行機器學習所得結果最佳，故將此關鍵變數取代原先的類別變數導入製程，其良率由46%提升至83.3%以上，達到降低產品失效率。

The global semiconductor manufacturing process is becoming more and more sophisticated, and the wafer process is getting smaller and smaller, which is a big challenge for wafer inspection equipment, and electron beam inspection equipment has attracted much attention in recent years. It can inspect tiny wafers and display defects through image stacking.
It is hard to manufacture Electrons Beam inspection machine, and yield is low. How to find out the key factors, reduce failure problems, and give customers high-quality products is the most urgent problem at present.
This study is collect 503 cases, use data mining to analysis maternal data of Deflection Image Control. Principal component analysis was by SPSS software to extract the most relevant and common category variables,and use three kinds of machine learning: logistic regression, decision tree and neural network to catch out the results, include binary classification metrics, the area under the ROC curve, AUC and high machine learning accuracy, after cross-validation, get two key category variables(Gain and Offset).
Results indicate that, combined two key category variableswith and six major variables set by the original R&D department, the result by machine learning is bestest than 232 category variables by production line, and import to process, The yield rate has increased from 46% to 83.3%. In the semiconductor industry, this is a major discovery to reduce product failure rates.

簡禎富、林昀萱、鄭仁傑，建構模糊決策樹及其在有交互作用之半導體資料之資料挖礦以提升良率之研究，品質學報，P193-210，2008年6月1日。
邱寬旭，類神經網路簡介，機電整合雜誌社，P58-62，2000年10月。
張斐章、張麗秋、黃浩倫，類神經網路理論與實務，東華書局股份有限公司，2003年9月。
施又銘，探討集成式分類樹模型於高維度資料特徵之選取-以電子束檢測設備為例，國立成功大學工業與資訊管理學系碩士論文，2002。
陳品潔，應用資料探勘技術於半導體業製程良率改善之研究，國立交通大學管理學院碩士論文，2003。
林嘉瑩，應用於半導體製程資料分析之正規化機器學習方法，國立交通大學電子研究所碩士論文，2016。
盧瑜芬，使用三種資料探勘演算法-類神經網路、邏輯斯迴歸及決策樹-預測乳癌患者存活情形之效能比較，國防醫學院公共衛生學研究所碩士論文，2006。
高仲仁，運用類神經網路進行隧道岩體分類，國立中央大學應用地質研究所碩士論文，2001。
Assunção, M. D., et al. (2015). Big Data computing and clouds: Trends and future directions. Journal of Parallel and Distributed Computing 79: 3-15.
Bousdekis, A., Papageorgiou, N., Magoutas, B., Apostolou, D., & Mentzas, G. (2017). A proactive event-driven decision model for joint equipment predictive maintenance and spare parts inventory optimization. Procedia Cirp 59: 184-189.
Cai, J., Luo, J., Wang, S., & Yang, S. (2018). Feature selection in machine learning: A new perspective. Neurocomputing 300: 70-79.
Ennett, C. M., M. Frize., & E. Charette. (2004). Improvement and automation of artificial neural networks to estimate medical outcomes. Medical engineering & physics 26(4): 321-328.
Fayyad, U., Piatetsky-Shapiro, G., & Smyth, P. (1996). From data mining to knowledge discovery in databases. AI magazine 17(3): 37-37.
Georganos, S., Grippa, T., Vanhuysse, S., Lennert, M., Shimoni, M., Kalogirou, S., & Wolff, E. (2018). Less is more: Optimizing classification performance through feature selection in a very-high-resolution remote sensing object-based urban application. GIScience & remote sensing 55(2): 221-242.
Hayashi, H., Oomura, M., Ihata, N., Shinkawa, A., Fan, F., & Li, J. (2009). Detection of critical defects with E-beam technology for development and monitoring of advanced NAND processes. 2009 IEEE/SEMI Advanced Semiconductor Manufacturing Conference, IEEE.
H. F. Yu., F. L. Huang., & C. J. Lin. (2011). Dual coordinate descent methods for logistic regression and maximum entropy models. Machine Learning 85(1-2): 41-75.
Jusman, Yessi., Ng, Siew. & Abu Osman, Noor Azuan. (2014). Investigation of CPD and HMDS sample preparation techniques for cervical cells in developing computer-aided screening system based on FE-SEM/EDX. The Scientific World Journal 2014.
Ledesma, S., Cerda, G., Avina, G., Hernández, D., & Torres, M. (2008). Feature selection using artificial neural networks. Mexican International Conference on Artificial Intelligence, Springer.
MacCallum, R. (1999). Psychology 820 course packet, OH: The Ohio State University Press.
M. H. Chiu., Y. R. Yu., & L. C. Hao. (2016). The use of facial micro-expression state and Tree-Forest Model for predicting conceptual-conflict based conceptual change. Chapter Title & Authors P184.
Meisburger, D., Spallas, J., Werder, K., & Muray, L. (2015). Proposed architecture of a multicolumn electron-beam wafer inspection system for high-volume manufacturing. Journal of Vac,uum Science & Technology B, Nanotechnology and Microelectronics: Materials, Processing, Measurement, and Phenomena 33(6): 06FN01.
Noh, P. J., Rohb, T. H. & Hana, I. (2005). Prognostic personal credit risk model considering censored information. Expert Systems with Applications 28(4): 753-762.
Ren, W., Liu, X., Hu, X., Luo, X., Ji, X., Xi, Q., Chou, K., Ebert, M., Ma, E. (2019). Multi-beam technology for defect inspection of wafer and mask. 35th European Mask and Lithography Conference (EMLC 2019), International Society for Optics and Photonics.
Sandborn, P. A. and C. Wilkinson. (2007). A maintenance planning and business case development model for the application of prognostics and health management (PHM) to electronic systems. Microelectronics Reliability 47(12):1889-1901.
Seiler, H. (1983). Secondary electron emission in the scanning electron microscope. Journal of Applied Physics 54(11): R1-R18.
Tang, Z. (1991). Time series forecasting using neural networks vs. Box-Jenkins methodology. Simulation 57(5): 303-310.
Waldrop, M. M. (2016). The chips are down for Moore’s law. Nature News 530(7589): 144.
Wang, S., Zhang, Y., Zhan, T., Phillips, P., Zhang, Y. D., Liu, G., Lu, S., Wu, X. (2016). Pathological brain detection by artificial intelligence in magnetic resonance imaging scanning (invited review). Progress in Electromagnetics Research 156:105-133.
Wei, Y., Zhen, H., & Wenmeng, T. (2011). The application of ReliefF algorithm for identifying CTQ in complex products. 2011 2nd IEEE International Conference on Emergency Management and Management Sciences, IEEE.
Wirth, R. and J. Hipp. (2000). CRISP-DM: Towards a standard process model for data mining. Proceedings of the 4th international conference on the practical applications of knowledge discovery and data mining, Springer-Verlag London, UK.
Witten, I. H., Frank, E., Hall, M. A., & Pal, C. J. (2016). The WEKA workbench. Online appendix for Data mining: practical machine learning tools and techniques. Morgan Kaufmann.
X. Fang, F. C. Yu, G. M. Yang, & Y. Y. Qu. (2019). Regression analysis with differential privacy preserving. IEEE access 7: 129353-129361.
Understanding Logistic Regression, Equiskill.com, Acailable:https://www.equiskill.com/understanding-logistic-regression/(accessed JULY 5, 2018)
Draper, B. A., et al. (2003). Recognizing faces with PCA and ICA. Computer vision and image understanding 91(1-2): 115-137.
Malhi, A. and R. X. Gao (2004). PCA-based feature selection scheme for machine defect classification. IEEE transactions on instrumentation and measurement 53(6): 1517-1525.