無論半導體、面板、太陽能等科技產業，生產效率與產品品質為顧客滿意之重要指標；當生產流程已被精實化，而難以精進時，產品之品質檢驗即扮演重要角色。為達到上述目的，最好的檢驗方式為全檢，即能夠獲得每一件產品的品質資訊，但因龐大的檢驗成本和無法滿足產品交期，而滯礙難行。有鑑於此，虛擬量測 (Virtual Metrology, VM)技術便發展用於解決上述問題，在不增加抽測樣本的條件下，利用生產機台的製程參數 (Process data)納入推估模型架構進行產品品質之推估，並達到線上且即時之品質監控。全自動虛擬量測 (Automatic Virtual Metrology, AVM)系統主要可分成三大部分，包含資料前處理，預測模型以及監控機制。監控機制又包含了製程參數相似度指標 (Similarity Index, SI)以及預測信心指標 (Reliance Index, RI)。然而目前AVM系統中信心指標發生異常時，表示預測值信心度不足，但未能提供問題發生之根本原因。因此，本論文將提出一精進信心指標機制異常分析方法，於信心指標發生異常時，解析AVM系統中之預測模型，即時判斷及辨識發生異常之製程參數，並通知設備工程師進行機台檢查，以確保整體製程品質與維護AVM系統預測精度。

Production efficiency and product quality are the most important features of customer satisfaction for all high-technology industries. For this reason, commodity inspection plays a significant role. “Virtual Metrology (VM)” concept is derived in order to meet the requirement of commodity inspection, and our research team has developed the automatic virtual metrology (AVM) system for various VM applications since 2007. VM makes conjecturing workpiece quality based on process data collected from production equipment with a slight supplement of actual metrology data feasible for semiconductor manufacturing.
AVM system consists of three modules, including data pre-process, conjecture and monitor scheme modules. Monitor scheme contains data quality index (DQI_X), similarity index (SIs), and reliance index (RI). Although many researches worked on the evaluation of prediction results such as confidence interval, confidence value, reliance value, and so on, few of these works aimed to find the reasons why the reliance index is under the threshold. This paper proposes a novel method to identify and determine the primary cause of an RI alarm in a virtual metrology system (VMS). The results will be sent to users and equipment engineers for them to adjust and check the equipment so as to resolve VMS manufacturability issues and ensure the production quality.
The illustrative examples involving semiconductor foundry are presented. Experimental results demonstrate that the proposed method is applicable to the VMS of production equipment (such as semiconductor and TFT-LCD).

[1] 2003 International Technology Roadmap for Semiconductors (ITRS), Dec. 2003. http://public.itrs.net/
[2] Y.-C. Chang and F.-T. Cheng, “Application Development of Virtual Metrology in Semiconductor Industry, in Proc. the 31st Annual Conference of the IEEE Industrial Electronics (IECON 2005), Raleigh, NC, Nov. 2005, pp. 124-129.
[3] P.-H. Chen, S. Wu, J.-S. Lin, F. Ko, H. Lo, J. Wang, C.-H. Yu, and M.-S. Liang, “Virtual Metrology: a Solution for Wafer to Wafer Advanced Process Control,” in Proc. 2005 IEEE International Symposium on Semiconductor Manufacturing, San Jose, CA, Sept. 2005, pp.155-157.
[4] Y.-C. Su, M.-H. Hung, F.-T. Cheng, and Y.-T. Chen, “A Processing Quality Prognostics Scheme for Plasma Sputtering in TFT-LCD Manufacturing,” IEEE Transactions on Semiconductor Manufacturing, vol. 19, no. 2, pp. 183-194, May 2006.
[5] Y.-C. Su, F.-T. Cheng, M.-H. Hung, and H.-C. Huang, “Intelligent Prognostics System Design and Implementation,” IEEE Transactions on Semiconductor Manufacturing, vol. 19, no. 2, pp. 195-207, May 2006.
[6] Y.-J. Chang, Y. Kang, C.-L. Hsu, C.-T. Chang, and T. Y. Chan, “Virtual Metrology Technique for Semiconductor Manufacturing,” in Proc. 2006 International Joint Conference on Neural Networks, July 2006, pp. 5289-5293.
[7] M.-H. Hung, T.-H. Lin, F.-T. Cheng, and R.-C. Lin, “A Novel Virtual Metrology Scheme for Predicting CVD Thickness in Semiconductor Manufacturing,” IEEE/ASME Transactions on Mechatronics, vol. 12, no. 3, pp. 308-316, June 2007.
[8] F.-T. Cheng, H.-C. Huang, and C.-A. Kao, “Dual-Phase Virtual Metrology Scheme,” IEEE Transactions on Semiconductor Manufacturing, vol. 20, no. 4, pp. 566-571, Nov. 2007
[9] 鄭芳田，全自動虛擬量測，科學發展月刊，第490期，Oct 2013。
[10] G. Chryssolouris, M. Lee, and A. Ramsey, “Confidence Interval Prediction for Neural Network Models,” IEEE Transactions on Neural Networks, vol. 7, no. 1, pp. 229-232, 1996.
[11] I. Rivals, and L. Personnaz, “Construction of Confidence Intervals for Neural Networks Based on Least Square Estimation,” Neural Networks, vol. 13, pp.463-484, 2000.
[12] D. Drurdjanovic, J. Lee, and J. Ni, “Watchdog Agent -- An Infotronics-Based Prognostics Approach for Product Performance Degradation Assessment and Prediction,” Advanced Engineering Informatics, vol. 17, pp.109-125, 2003.
[13] J. Yan and J. Lee, “Introduction of Watchdog Prognostics Agent and Its Application to Elevator Hoistway Performance Assessment,” Journal of the Chinese Institute of Industrial Engineers, vol. 22, no. 1, pp. 56-63, 2005.
[14] F.-T. Cheng, Y.-T. Chen, Y.-C. Su, and D.-L. Zeng, “Evaluating reliance level of a virtual metrology system,” IEEE Transactions on Semiconductor Manufacturing, vol. 21, no. 1, pp. 92-103, Feb. 2008.
[15] S. Wold, M. Sjöström, and L. Eriksson, “PLS-Regression: A Basic Tool of Chemometrics,” Chemometrics and Intelligent Laboratory Systems, vol. 58, issue 2, pp. 109-130, Oct. 2001.
[16] S. Haykin, Neural Networks, A Comprehensive Foundation, Prentice Hall, NJ:07458, 1999.
[17] Jerome H. Friedman and Werner Stuetzle, “Projection Pursuit Regression”, Journal of the American Statistical Association, Vol. 76, No. 376 (Dec., 1981), pp. 817- 823.
[18] Jeng-Neng Hwang, Shyh-Rong Lay. Martin, Maechler. R. Douglas Martin and Jamej Schimert, “Regression Modeling in Back-Propagation and Proiection Pursuit Learning”, IEEE Transaction on Neural Network. 1994, vol.5, issue 3, pp.342-353.
[19] Jellq-Nellg, Hwang, Hang Li, Martin Maechler, R. Douglas Martin, Jim Schimert, “A Comparison of Projection Pursuit and Neural Network Regression Modeling”, Advances in Neural Information Processing Systems 4, [NIPS Conference, Denver, Colorado, USA, December 2-5, 1991]
[20] 簡媜富、施義成、林振銘、陳瑞坤編著，半導體製造技術與管理，初版，國立清華大學出版社，民國94年5月，P4-1~P4-33
[21] 吳建偉，”應用6標準差手法於TFT-LCD ODF製程改善研究”，中華民國品質學會，2007