臺灣博碩士論文加值系統

標準工時對於建造智慧工廠時有其重要性，制定標準工時在智慧工廠中可
以優化資源分配、最佳化生產排程與生產效率的提升。
本研究透過制定標準工時掌握現場流程，再利用標準工時有效管理生產效
率、產能規劃與預測，通常搜集到的標準工時都會使用機器學習之手法進行預
測。因此本研究將使用個案公司現有工時數據建立預測模型，並將進行評估使
用不同方法，包括：多元線性迴歸、Lasso 迴歸、Ridge 迴歸、決策樹迴歸、隨
機森林、支援向量機、梯度提升、Catboost、類經網路等的效能。本研究將運用
以上手法為個案公司之金屬齒輪加工流程制定標準工時，蒐集齒輪研磨流程之
生產數據，建立預測模型。本研究除了考慮傳統的材料屬性、工具規格和操作
條件等變量外，研究中也考量了工時預測的特定參數。包含了模數、齒數、齒
厚、數量、連續加工平均時間、首三件平均加工時間和首三件平均單件時間。
這些參數特徵讓我們能夠更全面和精確地建立標準工時的預測模型。
本研究預期透過預測模型的開發，能使個案公司能夠全面掌握現場流程並
進行有效的管理，也能透過預測模型協助企業適應市場需求變化，做出適當的
生產策略與投入資源的調整。

In the construction of smart factories, standard processing time play a crucial role in optimizing resource allocation, production scheduling, and improving production efficiency. This study aims to establish standard processing time to grasp on-site processes and then use these hours to effectively manage production efficiency, capacity planning, and forecasting. Typically, standard work hours data are collected and used in machine learning models for prediction. Therefore, this study utilizes existing processing time data from a case company to build predictive models and evaluate the performance of different methods, including multiple linear regression,
Lasso regression, Ridge regression, Classification & regression tree, random forest, support vector machine, gradient boosting, CatBoost, and feedforward neural networks. This study will employ these methods to establish standard work hours for the metal gear processing workflow of the case company by collecting
production data from the gear grinding process and developing pr edictive models.
In addition to traditional variables such as material properties, tool specifications, and operating conditions, this study also considers specific parameters for work hours prediction. These parameters include module, number of teeth, too th thickness, quantity, average continuous processing time, average processing time for the first three pieces, and average single-piece processing time for the first three pieces. These feature parameters allow for a more comprehensive and accurate establishment of predictive models for standard work hours. This study expects that through the development of predictive models, the case company will be able to comprehensively grasp on-site processes and perform effective management.
Additionally, the predictive models can help the enterprise adapt to market demand changes and make appropriate production strategies and resource adjustments.

摘要.................................... i
Abstract ............................... ii
誌謝.....................................iv
目錄.....................................v
表目錄 ..................................vii
圖目錄 ..................................viii
第一章 緒論 .............................1
1.1 研究背景與動機 .......................1
1.2 研究問題與研究目的....................3
1.3 研究架構 ............................4
第二章 文獻探討 .........................7
2.1 標準工時相關研究 ......................7
2.2 機器學習應用 .........................9
2.3 多元線性迴歸之應用......................12
2.4 決策迴歸樹之應用 ....................... 12
2.5 隨機森林迴歸之應用....................... 13
2.6 支援向量機迴歸之應用 .................... 14
2.7 梯度提升之應用 ......................... 16
2.8 Catboost 之應用 ............................... 16
2.9 類神經網路之應用 ................................... 17
第三章 個案說明與案例探討 ..................................... 24
3.1 齒輪製造產業之產業背景說明 ...................................... 24
3.2 齒輪的種類與基本介紹 ..................................... 25
3.3 齒輪製程說明 ............................29
3.4 個案簡介與問題描述................................ 32
第四章 研究方法 .......................................... 33
4.1 研究流程 ............................ 33
4.2 資料準備與資料初步處理 ................................... 34
4.3 探索性數據分析 .................................................. 34
4.3.1 資料分佈狀態 .............................. 35
4.3.2 資料相關性分析 ................................... 35
4.4 特徵工程 .............................................................. 36
4.4.1 特徵選取 ........................................................ 37
4.4.2 資料標準化 .................................... 37
4.5 資料劃分與驗證 ...................................... 37
4.6 預測模型之建構 ............................................. 38
4.6.1 多元線性迴歸 .................................. 38
4.6.2 決策迴歸樹 ....................................39
4.6.3 隨機森林迴歸 ..................................................40
4.6.4 支援向量迴歸 ..................................................42
4.6.5 梯度提升 .....................................................44
4.6.6 Catboost ....................................................46
4.6.7 類神經網路 ...................................................47
4.7 模型評估指標 ...................................................49
第五章 實證分析 ....................................................51
5.1 模型最佳參數與結果分析...........................................51
5.1.2 多元線性迴歸模型結果評估與分析 .................................57
5.1.3 決策迴歸樹模型結果評估與分析 ...................................58
5.1.4 隨機森林迴歸模型結果評估與分析 .................................59
5.1.5 支援向量迴歸模型結果評估與分析 .................................59
5.1.6 梯度提升模型結果評估與分析 .....................................60
5.1.7 Catboost 模型結果評估與分析 .................................. 60
5.1.8 類神經網路模型結果評估與分析 ...................................60
5.2 綜合評估與分析 ..................................................61
第六章 結論與未來研究方向 ............................................63
6.1 研究結論 ........................................................63
6.2 未來研究方向 ....................................................63
參考文獻 ...........................................................62
Abstract ......................................................... 70

[1] 石伊蓓, 吳育仁, 馮展華, & 廖昆隆. (2020). 台灣齒輪設計與製造智慧現
況. 機械資訊月刊. https://www.tami.org.tw/sp1/print/761/761_01.html
[2] 李家岩, & 洪佑鑫. (2022). 製造數據科學: 邁向智慧製造與數位決策. 前
程文化事業股份有限公司.
https://books.google.com.tw/books?id=_tA_zwEACAAJ
[3] 林樹強 . (2017). 工作研究 : 工 業 4.0. 鼎茂圖書出版股份有限公司 .
https://books.google.com.tw/books?id=R5uIzwEACAAJ
[4] 邱國祥. (2020). 以多元線性迴歸與機器學習模型預估不動產價格 -以台中
市實價登錄為例 國立中興大學]. 臺灣博碩士論文知識加值系統. 台中市.
https://hdl.handle.net/11296/nd3u33
[5] 洪雪娟 , & 陳 怡 靜 . (2023). 金 屬 製 品 產 業 韌 性 供 應 . 工 業 材 料 雜 誌 .
https://www.materialsnet.com.tw/DocView.aspx?id=52176
[6] 陳 士 端 . (2006). 齒 輪 製 造 業 的 回 顧 、 現 況 與 展 望 . 機 械 資 訊 月 刊 .
http://www.tami.org.tw/print/588/588_01.htm
[7] 賀力行, 林淑萍, & 蔡明春. (2008). 統計學: 觀念, 方法, 應用. 前程文化.
https://books.google.com.tw/books?id=kmpgtwAACAAJ
[8] 楊琮文. (2022). 多元線性迴歸預測及分析嘉義市區細懸浮微粒 國立中正
大學]. 臺灣博碩士論文知識加值系統. 嘉義縣.
https://hdl.handle.net/11296/uxr9nm
[9] 趙泓諭. (2023). 使用機器學習與多元線性迴歸模型分析台灣股票匯率趨勢
（ 2018-2022 年 ) 逢 甲 大 學 ]. 臺 灣 博 碩 士 論 文 知 識 加 值 系 統 . 台中市 .
https://hdl.handle.net/11296/jx7ydz
[10] 劉德騏, 林清源, 張金隆, 張煌權, & 何寅綺. (2020). 高階齒輪加工機國產
化的推動. 機械工業雜誌, 444, 5-10.
[11] 齒輪 ABC 編寫組. (2006). 齒輪入門. 小原歯車工業株式会社.
[12] 戴辰. (2023). 台灣齒輪產業大未來與 ESG 永續經營系列三 齒輪產業未
來市場趨勢與機遇. 工商時報.
https://www.ctee.com.tw/news/20230704700442-431203
[13] 簡禎富 , & 許 嘉 裕 . (2014). 資 料 挖 礦 與 大 數 據 分 析 . 前 程 文 化 .
https://books.google.com.tw/books?id=zELurQEACAAJ
[14] Albon, C. (2018). Machine Learning with Python Cookbook: Practical
Solutions from Preprocessing to Deep Learning. O'Reilly Media.
74
https://books.google.com.tw/books?id=kIhQDwAAQBAJ
[15] Bekesiene, S., & Meidute-Kavaliauskiene, I. (2022). Artificial Neural Networks
for Modelling and Predicting Urban Air Pollutants: Case of Lithuania.
Sustainability, 14(4). https://doi.org/10.3390/su14042470
[16] Breiman, L. (2001). Random forests. Machine Learning, 45(1), 5-32.
https://doi.org/10.1023/a:1010933404324
[17] Çakıt, E., & Dağdeviren, M. (2023). Comparative analysis of machine learning
algorithms for predicting standard time in a manufacturing environment.
Artificial Intelligence for Engineering Design, Analysis and Manufacturing, 37.
https://doi.org/10.1017/s0890060422000245
[18] Chang, W., Wang, X., Yang, J., & Qin, T. (2023). An Improved CatBoost-Based
Classification Model for Ecological Suitability of Blueberries. Sensors, 23(4).
[19] Chiang, Y.-M., Chang, L.-C., & Chang, F.-J. (2004). Comparison of staticfeedforward and dynamic-feedback neural networks for rainfall–runoff
modeling. Journal of Hydrology, 290(3), 297-311.
https://doi.org/https://doi.org/10.1016/j.jhydrol.2003.12.033
[20] Chu, H., Dong, K., Yan, J., Li, Z., Liu, Z., Cheng, Q., & Zhang, C. (2023).
Flexible process planning based on predictive models for machining time and
energy consumption. The International Journal of Advanced Manufacturing
Technology, 128, 1-18. https://doi.org/10.1007/s00170-023-12027-3
[21] Cornelius, E., Akman, O., & Hrozencik, D. (2021). COVID-19 Mortality
Prediction Using Machine Learning-Integrated Random Forest Algorithm under
Varying Patient Frailty. Mathematics, 9(17).
https://doi.org/10.3390/math9172043
[22] Ersöz, T., Güven, I., & Ersöz, F. (2022). Defective products management in a
furniture production company: A data mining approach [Article]. Applied
Stochastic Models in Business and Industry , 38(5), 901-914.
https://doi.org/10.1002/asmb.2685
[23] Falkenberg, S. F., & Spinler, S. (2023). Integrating operational and human
factors to predict daily productivity of warehouse employees using extreme
gradient boosting. International Journal of Production Research, 61(24), 8654-
8673. https://doi.org/10.1080/00207543.2022.2159563
75
[24] Friedman, J. H. (2001). Greedy function approximation: A gradient boosting
machine. The Annals of Statistics, 29(5), 1189-1232, 1144.
https://doi.org/10.1214/aos/1013203451
[25] Gatarić, D., Ruškić, N., Aleksić, B., Đurić, T., Pezo, L., Lončar, B., & Pezo, M.
(2023). Predicting Road Traffic Accidents—Artificial Neural Network
Approach. Algorithms, 16(5).
[26] Géron, A. (2019). Hands-on Machine Learning with Scikit-Learn, Keras, and
TensorFlow. O’Reilly Media, Inc.
[27] Gómez, M. M. (2020). Prediction of work-related musculoskeletal discomfort
in the meat processing industry using statistical models [Article]. International
Journal of Industrial Ergonomics, 75, 9, Article 102876.
https://doi.org/10.1016/j.ergon.2019.102876
[28] Hastie, T., Tibshirani, R., & Friedman, J. H. (2009). The Elements of Statistical
Learning: Data Mining, Inference, and Prediction . Springer.
https://books.google.com.tw/books?id=eBSgoAEACAAJ
[29]Jain, K., & Choudhary, N. (2022). Comparative analysis of machine learning
techniques for predicting production capability of crop yield. International
Journal of System Assurance Engineering and Management, 13(S1), 583-593.
https://doi.org/10.1007/s13198-021-01543-8
[30] Kanal, L. (2003). Perceptron. 1383-1385.
[31] Kang, K.-S., Kim, T.-H., & Rhee, I.-K. (1994). The establishment of standard
time in die manufacturing process using standard data. Computers & Industrial
Engineering, 27(1), 539-542. https://doi.org/https://doi.org/10.1016/0360-
8352(94)90353-0
[32] Ko, C. S., Cha, M. S., & Rho, J. J. (2007). A case study for determining standard
time in a multi-pattern and short life-cycle production system. Computers &
Industrial Engineering, 53(2), 321-325.
https://doi.org/10.1016/j.cie.2007.06.025
[33] Koyama, K., Tanaka, M., Cho, B.-H., Yoshikawa, Y., & Koseki, S. (2021).
Predicting sensory evaluation of spinach freshness using machine learning
model and digital images. PLOS ONE, 16, e0248769.
https://doi.org/10.1371/journal.pone.0248769
76
[34] Kutschenreiter-Praszkiewicz, I. (2008). Application of artificial neural network
for determination of standard time in machining. Journal of Intelligent
Manufacturing, 19(2), 233-240. https://doi.org/10.1007/s10845-008-0076-6
[35] Liao, H.-y., Cade, W., & Behdad, S. (2021). Forecasting Repair and
Maintenance Services of Medical Devices Using Support Vector Machine.
Journal of Manufacturing Science and Engineering, 144(3).
https://doi.org/10.1115/1.4051886
[36] McGill, R., Tukey, J. W., & Larsen, W. A. (1978). Variations of Box Plots. The
American Statistician, 32(1), 12-16. https://doi.org/10.2307/2683468
[37] Monego, V. S., Anochi, J. A., & de Campos Velho, H. F. (2022). South America
Seasonal Precipitation Prediction by Gradient-Boosting Machine-Learning
Approach. Atmosphere, 13(2).
[38] Morapedi, T. D., & Obagbuwa, I. C. (2023). Air pollution particulate matter
(PM2.5) prediction in South African cities using machine learning techniques.
Front Artif Intell, 6, 1230087. https://doi.org/10.3389/frai.2023.1230087
[39] Pearson, K. (1900). X. On the criterion that a given system of deviations from
the probable in the case of a correlated system of variables is such that it can
be reasonably supposed to have arisen from random sampling. The London,
Edinburgh, and Dublin Philosophical Magazine and Journal of Science ,
50(302), 157-175. https://doi.org/10.1080/14786440009463897
[40] Prokhorenkova, L., Gusev, G., Vorobev, A., Dorogush, A. V., & Gulin, A. (2018).
CatBoost: unbiased boosting with categorical features Proceedings of the 32nd
International Conference on Neural Information Processing Systems, Montréal,
Canada.
[41] Rashid, T. (2016). Make Your Own Neural Network: A Gentle Journey Through
the Mathematics of Neural Networks, and Making Your Own Using the Python
Computer Language. CreateSpace Independent Publishing Platform.
https://books.google.com.tw/books?id=Zli_jwEACAAJ
[42] Rodrigues, A., Silva, F. J. G., Sousa, V. F. C., Pinto, A. G., Ferreira, L. P., &
Pereira, T. (2022). Using an Artificial Neural Network Approach to Predict
Machining Time. Metals, 12(10).
[43] Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986). Learning
77
representations by back-propagating errors. Nature, 323(6088), 533-536.
https://doi.org/10.1038/323533a0
[44] Sahadevan, D., Al Ali, H., Notman, D., & Mukandavire, Z. (2023). Optimising
Airport Ground Resource Allocation for Multiple Aircraft Using Machine
Learning-Based Arrival Time Prediction. Aerospace, 10(6).
https://doi.org/10.3390/aerospace10060509
[45] Singh, B. (2021). Predicting airline passengers’ loyalty using artificial neural
network theory. Journal of Air Transport Management, 94, 102080.
https://doi.org/https://doi.org/10.1016/j.jairtraman.2021.102080
[46] Swamynathan, M. (2017). Mastering Machine Learning with Python in Six
Steps. https://doi.org/10.1007/978-1-4842-2866-1
[47] Tewari, P. C. (2017). Work Study and Ergonomics. CRC Press.
https://books.google.com.tw/books?id=TR-qtAEACAAJ
[48] Wang, T., Hu, S. H., & Jiang, Y. (2021). Predicting shared-car use and
examining nonlinear effects using gradient boosting regression trees [Article].
International Journal of Sustainable Transportation, 15(12), 893-907.
https://doi.org/10.1080/15568318.2020.1827316
[49] Wu, D., Jennings, C., Terpenny, J., Gao, R. X., & Kumara, S. (2017). A
Comparative Study on Machine Learning Algorithms for Smart Manufacturing:
Tool Wear Prediction Using Random Forests. Journal of Manufacturing Science
and Engineering, 139(7). https://doi.org/10.1115/1.4036350
[50] Yan, X., Liu, X., & Zhao, X. (2020). Using machine learning for direct demand
modeling of ridesourcing services in Chicago. Journal of Transport Geography,
83. https://doi.org/10.1016/j.jtrangeo.2020.102661
[51] Yoon, J. (2020). Forecasting of Real GDP Growth Using Machine Learning
Models: Gradient Boosting and Random Forest Approach. Computational
Economics, 57(1), 247-265. https://doi.org/10.1007/s10614-020-10054-w
[52] Zhang, Q., Liu, F., Wan, X., & Xu, G. (2015). An Adaptive Support Vector
Regression Machine for the State Prognosis of Mechanical Systems. Shock and
Vibration, 2015, 469165. https://doi.org/10.1155/2015/469165