品質稽核是企業重要的活動，品質稽核結果進而影響客戶信心、公司營運和客戶訂單量；每年外部稽核活動有第三方定期稽核、客戶年度稽核、新客戶認證稽核和任務性稽核，這些稽核結果，皆會影響公司形象。
現在正處於工業 4.0資訊網路時代，資料資訊化已成為工作基本運作，許多工廠已導入機器人，以自動化生產和較低的勞動成本取代人力，增加企業產值與效能，並將此精神加入至品質稽核，導入智能化，快速找到企業流程和單位較有風險的地方。過往皆靠人員記憶的經驗和散布各地資料進行客戶來訪前的預防改善，資料無整理和進行標準化，往往方向錯誤和無法提前找出真正的問題，且需耗費許多人力成本進行資料分析和討論方向，最終結果並不穩定，常常導致客戶來訪缺失項目多和不滿意。
本研究利用自然語言方式，將過往稽核項目和結果進行結構化，將文字形資料依照各項規章制度、法規、作業標準和運作流程等範圍歸納出相對應標籤，透過平台系統和自然語言處理，建構一套模型和方法，以執行和完成資料預測，預測各廠區和各客戶來訪前預警項目和風險評估，將有效人力放置風險高的地方，有科學根據和邏輯進行稽核前的防守應對，讓品質稽核系統與公司智慧營運目標一致，系統診斷風險之智能化方式讓組織可以提前防守和改善風險較高的地方，有效提升客戶滿意度。

Quality audit is an important activity of the enterprise. The results of quality audit will affect the company's image, certification, customer confidence, and further affect the company's operations and the number of customer orders. The annual external audit activities include third-party regular audits, customer annual audits, new customer certification audits and task-based audits. In the past, the records of audit results relied on the experience of senior staff and fragmentary data. When the company conducts self-examination before the customer visit, such a result will lead to the wrong direction of preparation, failure to find out the real problem, and consume a lot of manpower for data analysis and discussion time. The above results will lead to many examination disadvantages and dissatisfied of the customers.
In the era of Industry 4.0, informatization has become a good tool to improve performance. In order to increase the output value and efficiency of enterprises, many factories introduce robots to replace human labor with automated production and lower labor costs. Intelligence is added to quality audit process to quickly find the more risky places in enterprise processes and units.
This study utilizes natural language processing to structure the audit items and results. The textual data summarizes the corresponding labels according to various laws and regulations, customer needs, etc., and constructs a set of models and methods. The prosposed method can detect the early warning items and risk assessment of each factory and each customer so that the organization can defend and improve the places with high risks in advance to enhance customer satisfaction.

摘要.............................2
ABSTRACT.........................3
誌謝.............................4
圖目錄...........................6
表目錄...........................7
第一章 緒論......................8
1.1 研究背景.....................8
1.2 研究動機....................10
1.3 研究目的....................11
1.4 研究架構....................11
第二章 文獻回顧探討..............13
2.1 品質管理系統................13
2.2 自然語言處理................18
2.2.1 源起與發展................18
2.2.2 自然語言處理的常見任務.....20
2.2.3 機器學習和深度學習.........23
2.3 Summary.....................26
第三章 研究方法..................27
3.1 資料前處理...................28
3.2 模型訓練.....................29
3.3 模型驗證.....................37
3.4 模型評估.....................38
第四章 實證分析...................41
4.1 公司簡介.....................41
4.2 現有痛點.....................41
4.3 資料前處理...................42
4.4 模型架構.....................45
4.5 模型表現.....................45
4.6 專案效益評估.................52
第五章 結論 .....................55
5.1 結論.........................55
5.2 未來研究與方向................55
參考文獻.........................57

Bre, F., Gimenez, J. M., & Fachinotti, V. D. (2018). Prediction of wind pressure coefficients on building surfaces using artificial neural networks. Energy and Buildings, 158, 1429-1441.

Bates, M. (1995). Models of natural language understanding. Proceedings of the National Academy of Sciences, 92(22), 9977-9982.

Boiral, O., Guillaumie, L., Heras‐Saizarbitoria, I., & Tayo Tene, C. V. (2018). Adoption and outcomes of ISO 14001: A systematic review. International Journal of Management Reviews, 20(2), 411-432.

Chung, J., Gulcehre, C., Cho, K., & Bengio, Y. (2014). Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv preprint arXiv:1412.3555.

Chollampatt, S., & Ng, H. T. (2018, April). A multilayer convolutional encoder-decoder neural network for grammatical error correction. In Proceedings of the AAAI conference on artificial intelligence (Vol. 32, No. 1).

Chowdhary, K. (2020). Natural language processing. Fundamentals of artificial intelligence, 603-649.

Chojnacka-Komorowska, A., & Kochaniec, S. (2019). Improving the quality control process using the PDCA cycle. Prace Naukowe Uniwersytetu Ekonomicznego we Wrocławiu, 63(4).

Chowdhury, G. G., & Chowdhury, S. (2003). Introduction to digital libraries. Facet publishing.

Devlin, J., Chang, M. W., Lee, K., & Toutanova, K. (2018). Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.

Dale, R. (2020). Natural language generation: The commercial state of the art in 2020. Natural Language Engineering, 26(4), 481-487.

Erauskin‐Tolosa, A., Zubeltzu‐Jaka, E., Heras‐Saizarbitoria, I., & Boiral, O. (2020). ISO 14001, EMAS and environmental performance: A meta‐analysis. Business Strategy and the Environment, 29(3), 1145-1159.

Fonseca, L. (2015). From quality gurus and TQM to ISO 9001: 2015: a review of several quality paths. International Journal for Quality Research (IJQR), 9(1), 167-180.

Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8), 1735-1780.

Jenab, K., Wu, C., & Moslehpour, S. (2018). Design for six sigma: A review. Management Science Letters, 8(1), 1-18.

Khurana, D., Koli, A., Khatter, K., & Singh, S. (2022). Natural language processing: State of the art, current trends and challenges. Multimedia Tools and Applications, 1-32.

Kobayashi, M., & Takeda, K. (2000). Information retrieval on the web. ACM Computing Surveys (CSUR), 32(2), 144-173.

Khashabi, D., Chaturvedi, S., Roth, M., Upadhyay, S., & Roth, D. (2018, June). Looking beyond the surface: A challenge set for reading comprehension over multiple sentences. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers) (pp. 252-262).

Kelleher, J. D., & Dobnik, S. (2018). What is not where: the challenge of integrating spatial representations into deep learning architectures. arXiv preprint arXiv:1807.08133.

Liu, P. J., Saleh, M., Pot, E., Goodrich, B., Sepassi, R., Kaiser, L., & Shazeer, N. (2018). Generating wikipedia by summarizing long sequences. arXiv preprint arXiv:1801.10198.

Minaee, S., Kalchbrenner, N., Cambria, E., Nikzad, N., Chenaghlu, M., & Gao, J. (2021). Deep learning--based text classification: a comprehensive review. ACM Computing Surveys (CSUR), 54(3), 1-40.

Mikolov, T., Chen, K., Corrado, G., & Dean, J. (2013). Efficient estimation of word representations in vector space. arXiv preprint arXiv:1301.3781.

Peters ME, Neumann M, Iyyer M et al. Deep contextualized word representations, 2018, https://arxiv.org/abs/1802.05365?source=post_pagef2a314075741

McCulloch, W. S., & Pitts, W. (1943). A logical calculus of the ideas immanent in nervous activity. The bulletin of mathematical biophysics, 5(4), 115-133.

Radford, A., Narasimhan, K., Salimans, T., & Sutskever, I. (2018). Improving language understanding by generative pre-training.

Ruder, S. (2016). An overview of gradient descent optimization algorithms. arXiv preprint arXiv:1609.04747.

Song, K., Tan, X., He, D., Lu, J., Qin, T., & Liu, T. Y. (2018). Double path networks for sequence to sequence learning. arXiv preprint arXiv:1806.04856.

Shezan, F. H., Lao, Y., Peng, M., Wang, X., Sun, M., & Li, P. (2022). NL2GDPR: Automatically develop GDPR compliant android application features from natural language. arXiv preprint arXiv:2208.13361.

Turing, A. M. (2009). Computing machinery and intelligence. In Parsing the turing test (pp. 23-65). Springer, Dordrecht.

Weizenbaum, J. (1966). ELIZA—a computer program for the study of natural language communication between man and machine. Communications of the ACM, 9(1), 36-45.

Wu, S., Liu, S., Sohn, S., Moon, S., Wi, C. I., Juhn, Y., & Liu, H. (2018). Modeling asynchronous event sequences with RNNs. Journal of biomedical informatics, 83, 167-177.

來新陽. (2021). ISO 稽核小組過程指引介紹. 品質月刊, 57(4), 12-14.

吳英志. (2019). 從 ISO 9001 品質管理系統邁向卓越經營模式. 品質月刊, 55(12), 5-9.

周靜幸. (2021). 敏捷內部稽核-風險與價值導向. 內部稽核, (113), 30-35.

陳翠媚, 郭修暐, & 李明賢. (2018). ISO 9001: 2015 [績效評估] 導入的因應對策之個案研究. 品質月刊, 54(10), 35-39.

蔡篤村. (2020). 大數據分析如何翻轉內部稽核. 內部稽核, (108), 32-38.

葉碧華. (2006). 製造業在 ISO9001: 2000 驗證稽核最常見品質管理缺失之分析.

劉睿承. (2020). 公司品質管理政策對經營績效之影響-以台灣輪胎製造業為例.

韓慧林, 王俊堯, & 張宏宇. (2017). ISO9001: 2015 改版實務. 危機管理學刊, 14(2), 11-22.

網路資源
ISO - International Organization for Standardization
iT 邦幫忙:https://ithelp.ithome.com.tw/
菜鳥學院 http://hk.noobyard.com/