在工業環境中，製造過程的穩定性對於確保產品品質至關重要，不穩定因素和異常模式可能導致加工缺陷和工件尺寸偏差，進而影響產品並增加成本。傳統的管制圖方法雖能監控製造過程並識別潛在的異常，但是傳統方法較耗時且監控成本較高。為了監控和管控製造過程中的產品品質，提出管制圖模式識別算法使用機器學習模型實現異常檢測和品質管制。因此，本論文提出結合AIC準則進行模型選擇的方法，再使用該準則所辨認的參數重新生成資料，提升支持向量機對非常態數據的辨識效能。經過蒙地卡羅模擬法及範例分析，結果證實所提出的方法能夠有效降低辨識錯誤率，且在非常態資料的管制圖下亦有良好的表現。 

    In industrial environments, the stability of manufacturing processes is crucial for ensuring product quality. Common causes and abnormal patterns in production can lead to processing defects and deviations in workpiece dimensions, affecting products and increasing costs. Traditional control chart methods, while capable of monitoring manufacturing processes and identifying potential anomalies, are time-consuming and costly to implement. Control chart pattern recognition algorithms use machine learning models for non-normality detection and quality control to monitor and control product quality during manufacturing. In this thesis, we combine the support vector machine method (SVM) with the Akaike information criterion for model selection to enhance the SVM's recognition performance for non-normal data. Monte Carlo simulations and case studies confirm that the proposed method effectively reduces recognition errors and performs well on control charts with non-normal data.

目錄
目錄	I
表目錄	III
圖目錄	IV
第一章	緒論	1
1.1	研究背景	1
1.2	研究目的	4
1.3	研究架構	5
第二章	文獻探討	6
2.1	傳統管制圖與管制圖圖形識別	6
2.1.1	傳統管制圖識別	6
2.1.2	管制圖模式介紹	7
2.2	支持向量機	9
2.2.1	線性支持向量機	9
2.2.2	非線性支持向量機	13
第三章	研究方法	14
3.1	研究流程	14
3.2	生成資料	15
3.2.1	蒙地卡羅	15
3.2.2	非常態分配	15
3.3	參數設置	16
3.3.1	管制圖樣式之參數設置	16
3.3.2	支持向量機之參數選擇	17
3.4	辨識效益評估	18
第四章	模擬結果與分析	19
4.1	模擬流程	19
4.2	支持向量機的識別結果	20
第五章	範例分析	25
5.1	範例資料之模式介紹	26
5.2	範例資料之識別結果	28
第六章	結論	29
參考文獻	30
表目錄
表2.1	SVM演算法之核函數	13
表4.1	模擬資料之參數設置	20
表4.2	N=30 時不同自由度與子群個數的辨識錯誤率	22
表4.3	N=50 時不同自由度與子群個數的辨識錯誤率	23
表4.4	N=100 時不同自由度與子群個數的辨識錯誤率	24
表5.1	實際資料之模式參數設置	27
表5.2	不同子群個數下的辨識錯誤率	28
圖目錄
圖2.1	六種典型管制圖模式	7
圖2.2	SVM分類之超平面示意圖	11
圖3.1	研究流程圖	14
圖4.1	N=30 時不同自由度與子群個數的辨識錯誤率直方圖	22
圖4.2	N=50 時不同自由度與子群個數的辨識錯誤率直方圖	23
圖4.3	N=100 時不同自由度與子群個數的辨識錯誤率直方圖	24
圖5.1	不同子群個數下的辨識錯誤率折線圖	28

[1]	Alcock, R. J., & Manolopoulos, Y. (1999). Time-series similarity queries employing a feature-based approach. 7th Hellenic conference on informatics, 
[2]	Cheng, C.-S., Chen, P.-W., & Ho, Y. (2022). Control Chart Concurrent Pattern Classification Using Multi-Label Convolutional Neural Networks. Applied Sciences, 12(2), 787. 
[3]	Cheng, C.-S., Chen, P.-W., Hsieh, Y.-C., & Wu, Y.-T. (2023). Multivariate Process Control Chart Pattern Classification Using Multi-Channel Deep Convolutional Neural Networks. Mathematics, 11(15), 3291. 
[4]	Cheng, C.-S., Ho, Y., & Chiu, T.-C. (2021). End-to-end control chart pattern classification using a 1D convolutional neural network and transfer learning. Processes, 9(9), 1484. 
[5]	Company, W. E. (1958). Statistical quality control handbook. AT & T Technologies. 
[6]	Cortes, C., & Vapnik, V. (1995). Support-vector networks. Machine learning, 20, 273-297. 
[7]	Duncan, A. J. (1986). Quality Control and Industrial Statistics, 5th. Ed. Irwin, Homewood, IL: Richard D Irvin. 
[8]	Fuqua, D., & Razzaghi, T. (2020). A cost-sensitive convolution neural network learning for control chart pattern recognition. Expert Systems with Applications, 150, 113275. 
[9]	Huang, J.-W., Lee, P.-J., & Jaysawal, B. P. (2021). Multiscale Control Chart Pattern Recognition Using Histogram-Based Representation of Value and Zero-Crossing Rate. IEEE Transactions on Industrial Electronics, 69(1), 684-693. 
[10]	Lee, P.-H., Torng, C.-C., Lin, C.-H., & Chou, C.-Y. (2022). Control chart pattern recognition using spectral clustering technique and support vector machine under gamma distribution. Computers & Industrial Engineering, 171, 108437. 
[11]	Li, Y., Ma, Y., Dai, W., & Zhang, W. (2023). Control Chart Pattern Recognition Based on MDWOP and Ensemble Classifier. 2023 IEEE International Conference on Industrial Engineering and Engineering Management (IEEM), 
[12]	Lu, C.-J., Shao, Y. E., & Li, C.-C. (2014). Recognition of concurrent control chart patterns by integrating ICA and SVM. Applied Mathematics & Information Sciences, 8(2), 681. 
[13]	Lu, Z., Wang, M., & Dai, W. (2020). A condition monitoring approach for machining process based on control chart pattern recognition with dynamically-sized observation windows. Computers & Industrial Engineering, 142, 106360. 
[14]	Miao, Z., & Yang, M. (2019). Control chart pattern recognition based on convolution neural network. Smart Innovations in Communication and Computational Sciences: Proceedings of ICSICCS 2017, Volume 2, 
[15]	Nelson, L. S. (1985). Interpreting Shewhart X control charts. Journal of Quality Technology, 17(2), 114-116. 
[16]	Ranaee, V., Ebrahimzadeh, A., & Ghaderi, R. (2010). Application of the PSO–SVM model for recognition of control chart patterns. ISA transactions, 49(4), 577-586. 
[17]	Smola, A. J., Schölkopf, B., & Müller, K.-R. (1998). The connection between regularization operators and support vector kernels. Neural networks, 11(4), 637-649. 
[18]	Tian, G., Liu, Y., Wang, N., Liang, X., & Sun, J. (2023). A Novel Deep Learning Approach to Online Quality Control for Multiple Time Series Process with Imbalanced Data. Available at SSRN 4633193. 
[19]	Ünlü, R. (2021a). Cost-oriented LSTM methods for possible expansion of control charting signals. Computers & Industrial Engineering, 154, 107163. 
[20]	Wang, F. K., Bizuneh, B., & Cheng, X. B. (2019). One‐sided control chart based on support vector machines with differential evolution algorithm. Quality and Reliability Engineering International, 35(6), 1634-1645. 
[21]	Wu, C., Hou, H., Lei, C., Zhang, P., & Du, Y. (2023). A Novel Scheme of Control Chart Patterns Recognition in Autocorrelated Processes. Mathematics, 11(16), 3589. 
[22]	Yu, Y., & Zhang, M. (2021). Control chart recognition based on the parallel model of CNN and LSTM with GA optimization. Expert Systems with Applications, 185, 115689. 
[23]	Zan, T., Chen, J., Sun, Z., Liu, Z., Wang, M., Gao, X., & Gao, P. (2024). Pattern recognition of control chart with variable chain length based on recurrent neural network.