在LED零組件組裝製程中，固晶和打線為影響製程不良率的關鍵因子，在固晶與打線機台運作過程中，工程師會根據以往經驗做參數的調整，並將產品透過人力做品質檢測，在尚未確認品質合格前，機台將呈現停滯狀態，使得調整參數及品質檢驗上缺乏效率。本研究利用某製造廠所提供之機台參數資料，根據資料調整前後之關係性，發展完善的長短期記憶遞迴神經網路(LSTM)預測模型，並結合工程師專業，挖掘適當預測模型之特徵，以提升整體模型預測率。
固晶資料集記錄了每次調機產生之固晶狀況及銀膠高度，並發現銀膠高度與固晶狀況的對應關係，透過決策樹將資訊整合，並利用分枝結果協助工程師做為調整參數之依據。
本研究之成果除運用在LED製程外，亦可應用此預測模式於相關產業的生產製程，藉此提升產線之生產效率。

During the assembly process of LED components, die bonding and wire bonding are the key factors affecting the defective rate. When the die bonding and wire bonding machine work, the engineers will adjust the parameters based on their experience. Meanwhile, the equipment wouldn’t be remained functioning until products passed through quality inspections. It is inefficient for tuning parameters and quality inspection. Therefore, the research utilizes the parameters of machine by a manufacturer. It dependents on their relationship and Constructs a Long Short-term Memory (LSTM) Recurrent Neural Network Prediction Model. During the process, the research also combines the information by engineers, trying to construct the new features for the prediction model. Finally, it improved the overall model of prediction rate.
In the die bonding data, the die bonding condition and the height of the Silver Epoxy were recorded on each parameter of the machine. The relationship between the height of the silver Epoxy and the status of the die bonding is observed. We integrated the information into the decision tree and its branches are used as well. The result assists the engineer to adjust the parameters.
The approach of this study can not only apply to the LED manufacturing process but also apply to similar area based on the same concept. It can improve the efficiency of a production line.

Amari, S.-I. (1998). "Natural gradient works efficiently in learning." Neural Computation 10(2)： 251-276.

Becker, S., and Y. Le Cun (1988). Improving the convergence of back-propagation learning with second-order methods. Proceedings of the 1988 connectionist models summer school, San Matteo, CA： Morgan Kaufmann.

Bengio, Y., et al. (1994). "Learning long-term dependencies with gradient descent is difficult." IEEE transactions on neural networks 5(2)： 157-166.

Deng, L., et al. (2013). Recent advances in deep learning for speech research at Microsoft. Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on, IEEE.

Ding, H., et al. (2008). "Querying and mining of time series data： experimental comparison of representations and distance measures." Proceedings of the VLDB Endowment 1(2)： 1542-1552.

Fu, T.-c. (2011). "A review on time series data mining." Engineering Applications of Artificial Intelligence 24(1)： 164-181.

Gers, F. A., et al. (1999). "Learning to forget： Continual prediction with LSTM."

Graves, A., et al. (2013). Speech recognition with deep recurrent neural networks. Acoustics, speech and signal processing (icassp), 2013 IEEE international conference on, IEEE.

Graves, A. and J. Schmidhuber (2005). "Framewise phoneme classification with bidirectional LSTM and other neural network architectures." Neural Networks 18(5-6)： 602-610.

Haq, A. A. U., et al. (2016). "Feature Construction for Dense Inline Data in Semiconductor Manufacturing Processes." IFAC-PapersOnLine 49(28)： 274-279.

Hochreiter, S. and J. Schmidhuber (1997). "Long short-term memory." Neural Computation 9(8)： 1735-1780.

Kingma, D. P. and J. Ba (2014). "Adam： A method for stochastic optimization." arXiv preprint arXiv：1412.6980.

Nurunnahar, S., et al. (2017). A short-term wind speed forecasting using SVR and BP-ANN： A comparative analysis. Computer and Information Technology (ICCIT), 2017 20th International Conference of, IEEE.

Petridis, S. and M. Pantic (2016). Deep complementary bottleneck features for visual speech recognition. Acoustics, Speech and Signal Processing (ICASSP), 2016 IEEE International Conference on, IEEE.

Qing, X. and Y. Niu (2018). "Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM." Energy 148： 461-468.

Rani, S. and G. Sikka (2012). "Recent techniques of clustering of time series data： a survey." International Journal of Computer Applications 52(15).

Rush, A. M., et al. (2015). "A neural attention model for abstractive sentence summarization." arXiv preprint arXiv：1509.00685.

Sundermeyer, M., et al. (2012). LSTM neural networks for language modeling. Thirteenth Annual Conference of the International Speech Communication Association.

Tieleman, T. and G. Hinton (2012). "Lecture 6.5-RMSProp, COURSERA： Neural networks for machine learning." University of Toronto, Technical Report.

Ting, J., et al. (2006). "Mining of stock data： intra-and inter-stock pattern associative classification." Threshold 5(100)： 5-99.

Viegas, R., et al. (2017). "Daily prediction of ICU readmissions using feature engineering and ensemble fuzzy modeling." Expert Systems with Applications 79： 244-253.

Wei, L. and E. Keogh (2006). Semi-supervised time series classification. Proceedings of the 12th ACM SIGKDD international conference on Knowledge discovery and data mining, ACM.

Williams, R. J. and D. Zipser (1989). "A learning algorithm for continually running fully recurrent neural networks." Neural Computation 1(2)： 270-280.

Xingjian, S., et al. (2015). Convolutional LSTM network： A machine learning approach for precipitation nowcasting. Advances in neural information processing systems.

Yang, K. and C. Shahabi (2005). On the stationarity of multivariate time series for correlation-based data analysis. Data Mining, Fifth IEEE International Conference on, IEEE.

Yang, Y., et al. (2018). "A CFCC-LSTM model for sea surface temperature prediction." IEEE Geoscience and Remote Sensing Letters 15(2)： 207-211.

Zdravevski, E., et al. (2015). Robust histogram-based feature engineering of time series data. Computer Science and Information Systems (FedCSIS), 2015 Federated Conference on, IEEE.

Zhao, Y., et al. (2018). "Applying deep bidirectional LSTM and mixture density network for basketball trajectory prediction." Optik-International Journal for Light and Electron Optics 158： 266-272.

張祐銜 and 劉正毓 (2009). 發光二極體的封裝技術, 科學發展.