在半導體封裝製程中，脫層(delamination)是影響產品良率的主要原因之一。脫層問題常會發生在元件的連接處，例如：晶粒(die)跟環氧模壓樹酯(epoxy molding compound, EMC)之間、環氧樹酯(epoxy)跟基板(substrate)之間、釘架(lead frame)跟環氧模壓樹酯(EMC)之間等地方。

本研究提出資料科學的架構，應用於半導體封裝製程的資料上，針對脫層問題及良率進行診斷。

首先，我們先以least absolute shrinkage and selection operator (LASSO) regression和stepwise regression進行變數挑選，從千以百計的參數中(包括機台型號、機台參數、製程、原料種類等)找出可能會造成脫層的主要因子，並與工程師商討進一步的工程驗證。接著，我們使用四個資料科學的模型：backpropagation neural network (BPN), support vector regression (SVR), partial least squares (PLS) and gradient boosting machine (GBM)，建構出脫層率的預測模型。在過程中，我們面臨到類別不平衡(class imbalance problem)的議題；而針對預測模型的結果，我們進一步探討其中的管理意涵，解析false positive 跟 false negative之間的議題，給予彈性決策的建議。最後，關於「何時該重新訓練模型」的議題，我們提出一個成本導向的方法給予建議。

實證研究是以台灣龍頭半導體封裝製造商所提供的資料進行，結果顯示在變數挑選階段，某些因子確實有機會造成脫層問題的發生，未來可以建構文件知識管理系統，提升管理品質；在預測模型階段，BPN跟GBM提供良好的準確率，並在false positive 跟 false negative的議題上提供彈性的決策方案。

In the semiconductor assembly process, delamination in die-attach layers is a leading cause of defective products. Delamination may exist between die and epoxy molding compound (EMC), epoxy and substrate, lead frame and EMC, etc. The troubleshooting is case-by-case and time-consuming without a systematic diagnosis approach. In this thesis, we propose a data science framework using least absolute shrinkage and selection operator (LASSO) regression and stepwise regression to extract the key variables affecting delamination. Then, we construct backpropagation neural network (BPN), support vector regression (SVR), partial least squares (PLS) and gradient boosting machine (GBM) to predict the ratio of the delamination area in a die. Besides, we investigate the imbalance between a false positive rate and false negative rate after the quality classification with BPN and GBM models to improve the tradeoff between the two types of risks. We validate the proposed framework with an empirical study of a semiconductor assembly company. The results show that the proposed framework provides delamination prediction with high accuracy and gains managerial insights for supporting the practical troubleshooting. Furthermore, since that the batch size determination of the dataset significantly affects the performance of the inline model retraining process, we suggest the cost-oriented method to address the issue.

Bishop, C. M. (2006). Pattern Recognition and Machine Learning (1st edition): Springer.
Chang, K.-H. (2015). Improving the efficiency and efficacy of stochastic trust-region response-surface method for simulation optimization. IEEE Transactions on Automatic Control, 60(5), 1235-1243.
Chien, C.-F., Hsu, C.-Y., & Chen, P.-L. (2013). Semiconductor fault detection and classification for yield enhancement and manufacturing intelligence. Flexible Services and Manufacturing Journal, 25(3), 367–388.
Chien, C.-F., & Chuang, S.-C. (2014). A framework for root cause detection of sub-batch processing system for semiconductor manufacturing big data analytics. IEEE Transactions on Semiconductor Manufacturing, 27(4), 475-488.
Chien, C.-F., Chang, K.-H., & Wang, W.-C. (2014). An empirical study of design-of-experiment data mining for yield-loss diagnosis for semiconductor manufacturing. Journal of Intelligent Manufacturing, 25(5), 961-972.
Chowdhury, A., Guenin, B., Woo, C., Kim, S., Lee, S. (1998). The effect of die attach layer delamination on the thermal performance of plastic packages. Electronic Components & Technology Conference 48th IEEE, 1140-1147.
Dong, Z.-H. (2017). Equipment Health Index for Fault Detection and Classification: An Empirical Study of Semiconductor Assembly Fabrication. M.S. thesis, Dept. IMIS, National Cheng Kung Univ., Tainan, Taiwan.
Duchi, J. & Singer, Y. (2009). Efficient online and batch learning using forwa rd backward splitting. Journal of Machine Learning Research, 10, 2873–2908.
Elkan, C. (2001). The foundations of cost-sensitive learning. Proceedings of the Seventeenth International Joint Conference on Artificial Intelligence, 2, 973-978.
Fausett, L. (1994). Fundamentals of Neural Networks: Architectures, Algorithms and Applications. Prentice-Hall, Inc.
Friedman, J., Hastie, T., & Tibshirani, R. (2001). The elements of statistical learning (Vol. 1): Springer series in statistics Springer, Berlin.
Han, J., Kamber, M., Pei, J. (2011). Data Mining: Concepts and Techniques (3rd edition): Morgan Kaufmann.
He, Q. Peter, & Wang, Jin. (2007). Fault Detection Using the k-Nearest Neighbor Rule for Semiconductor Manufacturing Processes. IEEE Transactions on Semiconductor Manufacturing, 20(4), 345-354.
Hecht-Nielsen, R. (1989). Theory of the backpropagation neural network. Proceedings of the International Joint Conference on Neural Networks (IJCNN), 593-605.
He, H., & Garcia, E.-A. (2009). Learning from imbalanced data. IEEE Transactions on Knowledge and Data Engineering, 21(9), 1263-1284.
Hocking, R. R. (1976). The analysis and selection of variables in linear regression. Biometrics, 32(1), 1-49.
Hsu, S.-C., & Chien, C.-F. (2007). Hybrid data mining approach for pattern extraction from wafer bin map to improve yield in semiconductor manufacturing. International Journal of Production Economics, 107(1), 88-103.
Hsu, C.-Y., Lin, S.-C., & Chien, C.-F. (2015). A back-propagation neural network with a distributed lag model for semiconductor vendor-managed inventory. Journal of Industrial and Production Engineering, 32(3), 149-161.
Kang, S., & Ramamohanarao K., K. (2014). A robust classifier for imbalanced datasets. In: Tseng V.S., Ho T.B., Zhou ZH., Chen A.L.P., Kao HY. (eds) Advances in Knowledge Discovery and Data Mining. Proceedings, Part I, 18th Pacific-Asia Conference, PAKDD 2014. Lecture Notes in Computer Science, vol. 8443.
Lee, C.-Y., & Chen, B.-S. (2017). Mutually-exclusive-and-collectively-exhaustive feature selection scheme. Applied Soft Computing (forthcoming).
Lee, C.-Y., & Tsai, T.-L. (2018). Data Science Framework for Yield Prediction and Control in TFT-LCD Manufacturing. Intelligent Data Analysis. (forthcoming).
Li, T.-S., Huang, C.-L., & Wu, Z.-Y. (2006). Data mining using genetic programming for construction of a semiconductor manufacturing yield rate prediction system. Journal of Intelligent Manufacturing, 17(3), 355-361
Liu, X.-Y., Wu, J., & Zhou, Z.-H. (2009). Exploratory undersampling for class-imbalance learning. IEEE Transactions on Systems, Man, and Cybernetics- Part B: Cybernetics, 39(2), 539-550.
MacGregor, JF, Jacckle, C, Kiparissdes, C, Kourti, M. (1994). Process monitoring and diagnosis by multiblock PLS methods. AIChE J, 40(5), 826–838.
Mcmahan, B. (2011). Follow-the-regularized-leader and mirror descent: Equivalence theorems and L1 regularization. Journal of Machine Learning Research, 15, 525–533.
Nomikos, P, & MacGregor, JF, (1994). Monitoring the processes using multiway principal component analysis. AIChE J, 40(8), 1361–1375.
Sauber, J., Lee, L., Hsu, S., Hongsmatip, T. (1994). Fracture properties of molding compound materials for IC plastic packaging. IEEE Transactions on Components, Packaging and Manufacturing Technology Society, 17(4), 533-541.
Shalev-Shwartz, S. (2011). Online learning and online convex optimization. Foundations and Trends in Machine Learning, 4(2), 107-194.
Shih, C.-Y. (2004). Study on Delamination of Bottom Paddles of Lead Frames for Semiconductor Packaging. M.S. thesis, Dept. ES, National Cheng Kung Univ., Tainan, Taiwan.
Smola, A., & Vapnik, V. (1997). Support vector regression machines. Advances in neural information processing systems, 9, 155-161.
Spitzlsperger, G, Schmidt, C, Ernst, G, Strasser, H, Speil, M. (2005). Fault detection for a via etch process using adaptive multivariate methods. IEEE Trans Semicond Manuf, 18(4), 528–533.
Stehman, V. S. (1997). Selecting and interpreting measures of thematic classification accuracy. Remote Sensing of Environment, 62(1), 77–89.
Tay, A. O., & Lin, T. (1996). Moisture Diffusion and Heat Transfer in Plastic IC Packages. IEEE Transactions on Components, Packaging, and Manufacturing Technology, 19(2), 186-193.
Tsai, T.-L. (2017). Data Science and Bayesian Decision in Extubation Prediction of Surgical Intensive Care Unit. M.S. thesis, Dept. IMIS, National Cheng Kung Univ., Tainan, Taiwan.
Tenenhaus, M., Vinzi, V. E., Chatelin, Y.-M., Lauro, C. (2005). PLS path modeling. Computational statistics & data analysis, 48(1), 159-205.
Tibshirani, R. (1996). Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society Series B (Methodological), 58(1), 267-288.
Yu, C.-M., Chien, C.-F., Kuo, C.-J. (2017). Exploit the value of production data to discover opportunities for saving power consumption of production tools. IEEE Transactions on Semiconductor Manufacturing, 30(4), 345-350.
Verdier, G., & Ferreira, A. (2011). Adaptive Mahalanobis Distance and k -Nearest Neighbor Rule for Fault Detection in Semiconductor Manufacturing. IEEE Transactions on Semiconductor Manufacturing, 24(1), 59-68.
Wikströma, C.Albano, C.Eriksson, L.Fride´n, H.Johansson, E, Nordahl, A, Ra¨nnar, S, Sandberg, M, Kettaneh-Wold, N, Wold, S. (1998). Multivariate process and quality monitoring applied to an electrolysis process part I. Process supervision with multivariate control charts. Chemom Intell Lab Syst, 42(1–2), 221–231.
Wise, BM, & Gallagher, NB. (1996). The process chemometrics approach to process monitoring and fault detection. J Process Contr, 6(6), 329–348.
Xiao, L. (2010). Dual averaging methods for regularized stochastic learning and online optimization. Journal of Machine Learning Research, 11, 2543-2596.
Zou, H. (2006). The adaptive lasso and its oracle properties. Journal of the American Statistical Association, 101(476), 1418-1429.