當產品在顧客的使用過程中，若在保固期內發生異常則會退回以釐清問題並排除之。本研究乃是針對顧客所退回之異常產品進行責任歸屬問題之判定，提出可能的改善對策。透過資料分析的方法，從現有客訴的產品資訊，取得相關之特性變數資料，本研究使用馬氏距離的分類方法，協助進行良品與不良品質判別。經由實際案例的分析，使用本研究所建議的方法可使判別正確率提升至17%左右，並能有效節省公司的作業成本支出。本研究之成果可提供產業界有關產品品質問題的分類上之參考。

When the product is used by the customer with a malfunction occurs during the warranty period, it will be returned to clarify the problem and figure out the solution.
This research is to determine the attribution of responsibility for abnormal products returned by customers and propose possible improvement method. Through the method of data analysis, the relevant characteristic variable data is obtained from the product information of the existing customer complaints.
This study uses the classification method of Mahalanobis distance to assist in the judgment of good and bad quality. Through the analysis of this actual cases, the use of the method suggested by this research can improve the accuracy of discrimination by approximately 17% and can effectively save the company's operating costs. The results of this research can provide a reference for the classification of product quality issues in the industry.

目錄
誌謝 I
摘要 II
ABSTRACT III
目錄 IV
第一章 緒論 1
1.1 研究背景 1
1.2 研究動機 2
1.3 研究目的 3
1.4 論文架構 4
第二章 文獻探討 5
2.1 電池與電池模組特性 5
2.2 馬氏距離 7
2.2.1 馬氏距離 7
2.2.2 馬氏距離的應用 11
2.3 MTS方法 12
第三章 研究方法 14
3.1 研究方法 14
3.2評估指標 17
第四章 個案分析 19
4.1個案背景 19
4.2問題定義 22
4.3 重要變數之篩選 25
4.4訓練及測試資料之建立 27
4.5量測尺度之構建與確認 28
4.6模式之驗證 34
4.7效益分析 37
第五章 結論 38
參考文獻 39

參考文獻
1. Asada, M. (2001). Wafer Yield Prediction by the Mahalanobis-Taguchi System. IIE Transactions, 14, 25-28.
2. Chung, J.-P. and Hong, H.-Y. (2021) ,Blade fault diagnosis using Mahalanobsis distance. Journal of Mechanical Science and Technology, 35,1377-1385.
3. Ji, H. (2021) Statistics Mahalanobis distance for incipient sensor fault detection and diagnosis. Chemical Engineering Science, 230, 1-10.
4. Hsiao, Y.-H. and Su, C.-T. (2009) Multi-class MTS for Saxophone Timbre Quality Inspection Using Waveform Shape-based Features. IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics, 39, 690-704.
5. Mohsen, K., Aida, M., and Tan, Z. (2019) O-ADPI: Online, Adaptive Deep-Packet Inspector Using Mahalanobis Distance Map for Web Service Attacks Classification, IEEE Access, 7, 1-16.
6. Su, C.-T. (2013), Quality Engineering: Off-Line Methods and Applications, CRC Press/Taylor & Francis Group.
7. Su, C.-T. (2017), Mahalanobis–Taguchi System and Its Medical Applications, Neuropsychiatry, 7(4), 316-320.
8. Su, C.-T. and Hsiao, Y.-H. (2007) An Evaluation of the Robustness of MTS for Imbalanced Data. IEEE Transactions on Knowledge and Data Engineering, 19(10), 1321-1332.
9. Su, C.-T. and Wang, H.-C. (2004) Robust Design of Credit Scoring System by the Mahalanobis-Taguchi System. Asian Journal on Quality, 5, 1-16.
10. Watabe, A., Komiya, K., Usuki, J., Suzuki, K., and Ikeda, H. (2005), Effective Designation of Specific Shots on Video Service System Utilizing Mahalanobis Distance, IEEE Transactions on Consumer Electronics, 51(1), 152-159.