跟隨資訊時代的快速發展的腳步，越來越多的網路產業已經物流產業萌生發展，在打包合箱方面必不可少的就是需要更大量的膠佈，然而更多的傳統小型工廠很能負荷時代的腳步，慢慢面臨衰退。最直接的原因是小型公司沒有辦法投入大量的人力物力資源，需要通過資訊化工廠提高直接的生產水準。
本研究通過對塗佈機制程進行分析，運用機器學習Support Vector Machine（SVM）支持向量機中的一類支持向量機One-Class Support Victor Machine(OCSVM),並且使用支援向量資料描述方法support vector data description（SVDD）進行異類點檢測和分類。SVDD 是一種重要的資料描述方法, 它能夠對目標資料集進行超球形描述, 並可用於異類點檢測或分類. 在現實生活中目標資料集通常包含多個樣本類, 且需要同時對每一個樣本類進行超球形描述。
研究的結果將用於檢測塗佈機制程中的異類點，便於更早的發現制程問題以及機器停擺時間點，可輔助分析影響原因，以達到利潤、產值最大化。

Follow the pace of the rapid development of the information age, more and more has the logistics industry initiation network industry, essential in terms of packaging or box is the need to be more a lot of tape, however more traditional small factory can load the pace of The Times, slowly facing recession.
The most direct reason is that small firms can't invest a lot of manpower and resources, need to improve the level of direct production through information chemical plant.
This study through the analysis of mechanism of coating process, using Machine learning Support Vector Machine (SVM) is One of the Support Vector Machine (SVM) in Support Vector Machine (SVM) One - Class Support Victor Machine (OCSVM) and Support Vector data description method is used to Support Vector data description (SVDD) heterogeneous point detection and classification.
SVDD is a kind of important method to describe the data, it is able to super spherical description of target data set, and can be used in heterogeneous point detection or classification. In real life target sample data sets usually contain more than one class, and at the same time for each spherical sample class to describe.
Research results will be used to detect different point in the mechanism of coating process, to facilitate earlier discovery process and machine lockout point in time, can assist this paper analyzes the reasons of influence, in order to achieve profits and value maximization.

中文文獻
[1]. 王國柱、劉建昌、李元（2015），稀疏性SVDD方法再故障檢測中的應用研究，東北大學學報，36(6)，761-764。
[2]. 王建林、馬琳鈺、邱科鵬、劉偉旻、趙利強 (2017)，基於svdd的多時段間歇過程故障檢測.，儀器儀錶學報，38(11)， 2752-2761。
[3]. 王輝 （2012），基於代表示例選擇與 SVDD 的多示例學習演算法研究，杭州電子科技大學，碩士論文
[4]. 李冠男、胡云鹏、陈焕新、黎浩荣、李炅、胡文举 (2015)， 基于svdd的冷水机组传感器故障检测及效率分析，化工学报，05，196-201。
[5]. 李琳 （2016），基於OCSVM的工業控制系統入侵偵測演算法研究，瀋陽理工大學，碩士論文
[6]. 吳康寧 （2017）,基於人工智慧下的機器學習歷史及展望研究，科技尚品，06。
[7]. 陳娜 （2019），走進德國近距離感受工業4.0，科教導刊，31，257-257。
[8]. 鐘志旺、陳建譯、唐濤、徐田華、王峰 (2018)，基於SVDD的道岔故障檢測和健康評估方法，西南交通大學學報，53(004)，842-849。
[9]. 陶新民、李晨曦、李青、任超、劉銳、鄒俊榮 (2019)，不均衡最大軟間隔svdd軸承故障檢測模型，振動工程學報，032(004)，718-729.
[10]. 於進、钱锋（2010），基於粒子群優化的高斯核函數聚類演算法，计算机工程，36（014），22-23。
英文文獻
[11]. Wang P.L, Ge Z.Q, Song Z.H (2009),Online fault monitoring for batch processes based on adaptive multimodel,Chinese Journal of Scientific Instrument,30,1347-1352.
[12]. Banerjee A, Burlina P, Diehl C. A (2006), support vector method for anomaly detection in hyperspectral imagery. IEEE Trans.on Geoscience and Remote Sensing , 2006, 44(8): 2282-2291.
[13]. Bernhard Schölkopf (2001),Estimating the Support of a High-Dimensional Distribution ,Neural Computation,13(7),1443-1471.
[14]. Brereton R G, Lloyd G R.(2010),Support vector machines for classifiation anessio, 135 (2): 230-267.
[15]. D A Zakoldaev, A V Shukalov1, I O Zharinov1 and O O Zharinov (2010), Modernization stages of the Industry 3.0 company and projection route for the Industry 4.0 virtual factory,Materials Science and Engineering,537,032-005.
[16]. David M.J. Tax & Robert P.W. Duin,(2004),Support vector data description, Machine Learning,54,45-66 .
[17]. Khediri, I. B. , Weihs, C. , & Limam, M. (2012). Kernel k-means clustering based local support vector domain description fault detection of multimodal processes. Expert Systems with Applications, 39(2), 2166-2171.
[18]. Liu X Q, Li K, Marion McAfee, George William Irwin (2011),Improved nonlinear PCA for process monitoring using support vector data description,Journal of Process Control ,21(9),1306-1317.
[19]. Paul R.Cohen & Edward A.Feigenbaum (2014) The Handbook of Artificial Intelligence (volume 3), Butterworth-Heinemann.
[20]. Tax D M J. & Muller K-R (2004) A consistency-based model selectior for one-class classification, IEEE, 1051-4651.
[21]. Vapnik, V. (1998) Statistical Learning Theory, Wiley-Interscience.
[22]. Wu Ding-Hai, Zhang Pei-Lin, Ren Guo-Quan (2011) Review of One-class Classification Method Based on Support Vector, Computer Engineering,37(5),187-189.
[23]. Chen K-Y, Jack Cheng, Chen W-W, Wang Q(2020)Data-driven predictive maintenance planning framework for mep components based on bim and iot using machine learning algorithms, Automation in Construction, 112,103-087.
[24]. Peng X, Du R, Zhang Z B (2019)Predicting pipeline leakage in petrochemical system through GAN and LSTM, Knowledge-Based Systems,175,50-61.