全自動虛擬量測在半導體產業有相當廣泛的應用，可將離線且具延遲特性之品質抽檢改成線上且即時之品質全檢。晶圓切割製程在整個製程結束後才會進行整批晶圓的全檢，若在製程當中發生問題，仍需等到出貨前全檢站點才可發覺，如此將可能產生大量缺陷品。晶圓切割製程導入全自動虛擬量測後，在發現製程異常時就可立即進行即時改善，如此就可避免後續整批晶圓的浪費。然而全自動虛擬量測應用於晶圓切割製程時，需進行晶圓崩缺(Wafer Chipping)的數值預測；但並非每片晶圓都有晶圓崩缺產生，且全自動虛擬量測系統也無法分辨晶圓有無崩缺產生，進而對於每片晶圓都給予預測值，造成使用者無法藉由全自動虛擬量測系統區分晶圓是否發生崩缺。
為解決上述問題，將晶圓切割品質監控分為兩階段，階段一開發以集成學習為基礎的自動分類機制(Automated Classification Scheme, ACS)預先判斷該片晶圓否有崩缺發生，若有崩缺發生再進行階段二之全自動虛擬量測，進行崩缺數值預測，期達到晶圓切割製程階段的線上即時品質監控。

Automatic Virtual Metrology (AVM) has a wide range of applications in the semiconductor industry. It can convert sampling inspection with metrology delay into real-time and online total inspection. For the current wafer sawing process, the whole lot of wafers is inspected at the end of entire process; therefore, a defect occurs during processing will only be detected until the process finish, which is too late and may cause massive defects. After implementing AVM to the Wafer Sawing process, when an abnormality is found, it can be improved immediately to avoid generating defects in the subsequent wafers. Therefore, there is a need to predict wafer-chipping occurrence before applying AVM to the wafer sawing process. However, chipping won’t happen to all wafers. The AVM system can predict the chipping value for each wafer when a chipping exists while AVM cannot distinguish if the wafer has chipping or not; in other words, users can’t differentiate whether the wafer is chipped through the AVM system.
To solve the above mentioned problem, the wafer sawing quality monitoring is divided into two stages. An Automated Classification Scheme (ACS) based on ensemble learning is developed in Stage 1 to pre-determine whether there is chipping in the wafer. If chipping is detected, then proceed to Stage 2 for the AVM system to predict the chipping value.

[1] 林秀玲，半導體封裝製程參數設計之研究，逢甲大學工業工程研究所，碩士論文，民國89年。
[2] 林志良，晶圓切割製程的穩健設計-六標準差與田口實驗設計的應用，國立高雄應用科技大學工業工程與管理系碩士班，碩士論文，民國98年。
[3] C. Li, C. Chang and M. Jeng, "Applying Regional Level-Set Formulation to Post sawing Four-Element LED Wafer Inspection," in IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), vol. 41, no. 6, pp. 842-853, Nov. 2011.
[4] C.-F. Chang, J.-L. Wu, Y.-C. Wang, “A Hybrid Defect Detection Method for Wafer level Chip Scale Package Images“, International Journal on Computer, Consumer and Control (IJ3C), Vol. 2, No.2,pp.25-36, 2013.
[5] T.-J. Su, Y.-F. Chen, J.-C. Cheng, C.-L. Chiu, “Automated image analysis for evaluation of wafer backside chipping“, The International Journal of Advanced Manufacturing Technology, Vol. 99, pp.2015–2023, 2018 .
[6] T.-J. Su, Y.-F. Chen, J.-C. Cheng, C.-L. Chiu, “An artificial neural network approach for wafer dicing saw quality prediction“, Microelectronics Reliability, vol.91, pp. 257–261, 2018.
[7] M. Vagues, “Analysing backside chipping issues of the die at wafer saw,” San Jose State University, San Jose, CA, 2003, retrieved November 23, 2011.
[8] H. He, E.-A. Garcia, “Learning from Imbalanced Data“, IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, vol. 21, no. 9, pp. 1263–1284, Sep. 2009.
[9] T.-G. Dietterich, “Ensemble Methods in Machine Learning,” Proc. Conf. Multiple Classifier Systems, pp. 1-15, 2000.
[10] F. Cheng, H. Huang and C. Kao, "Developing an Automatic Virtual Metrology System," in IEEE Transactions on Automation Science and Engineering, vol. 9, no. 1, pp. 181-188, Jan. 2012.
[11] D. Zhang, L. Qian, B. Mao, C. Huang, B. Huang and Y. Si, "A Data-Driven Design for Fault Detection of Wind Turbines Using Random Forest and XGboost," in IEEE Access, vol. 6, pp. 21020-21031, 2018.
[12] M. PAL, "Applying Regional Level-Set Formulation to Postsawing Four-Element LED Wafer Inspection," International Journal of Remote Sensing, Vol. 26, No. 1, pp.217–222, 10 January 2005.
[13] Z. Masetica, A. Subasib, “Congestive heart failure detection using random Forest classifier,” Compute Methods Programs Biomed,Vol.130 , pp. 54-64, 2016.
[14] S. A. Shevchik, F. Saeidi, B. Meylan and K. Wasmer, "Prediction of Failure in Lubricated Surfaces Using Acoustic Time–Frequency Features and Random Forest Algorithm," in IEEE Transactions on Industrial Informatics, vol. 13, no. 4, pp. 1541-1553, Aug. 2017.
[15] L. Torlay, M. Perrone-Bertolotti, E. Thomas, M. Baciu, “Machine learning–XGBoost analysis of language networks to classify patients with epilepsy," Brain Informatics, vol.4, pp. 159–169, 2017.
[16] W. Chen, K. Fu, J. Zuo, X. Zheng, T. Huang and W. Ren, "Radar emitter classification for large data set based on weighted-xgboost," in IET Radar, Sonar & Navigation, vol. 11, no. 8, pp. 1203-1207, 8 2017.
[17] J. Zhong, Y. Sun, W. Peng, M. Xie, J. Yang and X. Tang, "XGBFEMF: An XGBoost-Based Framework for Essential Protein Prediction," in IEEE Transactions on NanoBioscience, vol. 17, no. 3, pp. 243-250, July 2018.
[18] J. Bergstra, Y. Bengio, “Random search for hyper-parameter optimization”, J. Mach. Learn. Res., vol. 13, no. 1, pp. 281-305, 2012.
[19] L. Rokach, O. Maimon, Data Mining with Decision Trees:Theory and Applications, World Scientific Publishing Co. Pte.Ltd., 2008.