特徵選擇與決策理論在紗線分級之研究

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：37

、訪客IP：18.188.101.251

姓名

連信仲(Hsin-Chung Lien) 查詢紙本館藏

畢業系所

機械工程學系

論文名稱

特徵選擇與決策理論在紗線分級之研究
(Combining Feature Selection with Decision Theory for Yarn Grading)

相關論文

★ 使用實驗計劃法求得印刷電路板微鑽針最佳鑽孔參數	★ 滾針軸承保持架用材料之電鍍氫脆研究
★ 強制氧化及熱機處理對鎂合金AZ91D固相回收製程之研究	★ 滾針軸承保持架圓角修正之有限元素分析
★ 透過乾式蝕刻製作新型鍺全包覆式閘極電晶體元件	★ 窗型球柵陣列構裝翹曲及熱應力分析
★ 冷軋延對ZK60擠製材的拉伸與疲勞性質之影響	★ 熱引伸輔助超塑成形製作機翼整流罩之設計及分析
★ 超塑性鋁合金5083用於機翼前緣整流罩之研究	★ 輕合金輪圈疲勞測試與分析
★ 滾針軸承保持架之有限元分析	★ 鎂合金之晶粒細化與超塑性研究
★ 平板式固態氧化物燃料電池穩態熱應力分析	★ 固態氧化物燃料電池連接板電漿鍍膜特性研究
★ 7XXX系鋁合金添加Sc之顯微組織與機械性質研究	★ 高延性鎂合金之特性及成形性研究

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

[檢視]

[下載]

本電子論文使用權限為同意立即開放。
已達開放權限電子全文僅授權使用者為學術研究之目的，進行個人非營利性質之檢索、閱讀、列印。
請遵守中華民國著作權法之相關規定，切勿任意重製、散佈、改作、轉貼、播送，以免觸法。

摘要(中)

紗線等級評估是市場評估紗線價值的一個重要指標，本文利用紗線特徵數據，經由特徵選擇、圖形識別及倒傳遞神經網路等理論進行紗線等級評估。實驗過程是藉由學習樣本之紗線強度差異(Single yarn strength cv%),100公尺重量變異(100m weight cv%),紗線的均勻度變異(Yarn evenness cv%),黑板棉粒數(black-board neps no/g),紗線斷裂強度(Single yarn break-strength gf/tex)及100公尺重量公差(100 m weight tolerance %)等六項紗線特徵，作為6個輸入之紗線屬性向量(property vectors, PVs)，經屬性向量輸入至決策理論以評估紗線之客觀等級。
本文首先利用圖形識別理論尋求簡單有效之紗線等級評估系統。在紗線分級的學習過程中，經由Karhunen-Loeve (K-L) expansion方法，將紗線屬性向量轉換至單位正交主軸向量(principal axis vectors, PAVs)上，達成特徵選取與降階之目的，應用Bayes Classifier方法及最小距離方法，求取評估紗線等級之決策函數，並在不降低辨識率之前提下重複上述學習過程，尋找可辨識最低維度主軸向量之轉換矩陣及決策函數。測試過程中，決策函數藉主軸向量進行紗線等級評估，藉由實驗結果證實，本方法辨識向量數目將減少到只需一維主軸向量，即可進行快速而有效的紗線等級評估，達到簡化辨識系統之目的，並進而獲得公正客觀之紗線等級評估結果。
在倒傳遞神經網路結合圖形識別理論尋求精確之紗線等級評估系統。辨識結果顯示將紗線之六項屬性向量作為輸入向量，以倒傳遞神經網路進行紗線等級評估，可獲得客觀且精確之評估結果，為提升系統性能，本文更以倒傳遞神經網路結合K-L expansion方法進行紗線分級，經由實驗結果證實，K-L expansion方法將有效降低輸入向量之亂度與維度，使倒傳遞神經網路簡化至只需一維主軸向量作為辨識向量，即可達到辨識之效能，而當倒傳遞神經網路以四維主軸向量作為辨識向量時，則紗線等級評估系統將具有較佳之強健性及再學習能力，並可達到精確之評估結果。
最後本文藉由計算紗線族群有效距離(effective distance between clusters EDC)之方法作為特徵選擇( Feature Selection )之依據，以達到提供明顯而有效之紗線分級特徵，並進而提升紗線分級之作業效益。實驗結果顯示EDC方法對主軸向量(principal axis vectors PAVs)作特徵選擇時，其平均特徵誤排數目及平均誤排總距離分別只有K-L expansion方法之33.3%及16.7%，而且EDC方法亦可直接應用於屬性向量之特徵選擇上，在不降低辨識精度下可減少PVs量測項目，達到相對於前文之紗線分級方法中，分別在計算PAVs之時間上及量測PVs之項目上同時提升16.7%之效益。

摘要(英)

The present research investigates a simple and effective system for grading textile yarns by pattern recognition theory. During the learning processes of textile yarn grading, yarn property vectors (PVs) can be transferred to the principal axis vector (PAV) of the orthonormal function with Karhunen-Loeve (K-L) expansion, so as to select features and then reduce dimensions. Whereas, through the Bayes Classifier (BC) or the Minimum Distance (MD) method, the decision function for evaluating textile yarns grading can be obtained. Then under the premise of not sacrificing identification rate, the learning process can be repeated to search for the transfer matrix and the decision function with PAVs of the lowest identifiable dimension. The present system is shown to create a quick and effective grading system for textile yarns by using PAVs of only one dimension, thus both simplifying the identification system and providing objective grading results.
The grade of textile yarns is an important index in evaluating the yarn’s market value. This paper uses the backpropagation neural network (BNN) and Karhunen-Loeve (K-L) expansion method to construct a new and highly accurate grading system. Outcomes show that a highly accurate and neutral grading system can be obtained if the BNN learning sample is comprehensive or by adopting BNN with relearning technique (Self-Healing). Considering the possibility of reducing the dimension of BNN input vectors without losing the accuracy, this paper preprocesses the BNN grading system using the K-L expansion. Experiments demonstrate that the K-L expansion provides a way to reduce the input dimensions, and that a single principle axis value of BNN with K-L expansion grading system is able to grade textile yarns. In addition, the experiment demonstrates that as the input dimensions are reduced to four in a self-healing neural network with K-L expansion, the grading system provides the high accuracy and robustness.
By using the effective distance between clusters (EDC) as the basis for feature selection, this paper achieves a significant and effective feature for textile yarn grading, and further upgrades the operational efficiency of such grading. The results, such as the feature selection processing to principal axis vectors (PAVs) by EDC, show that The features’ average number and average total distance of mistaken ranking by EDC were only 33.3% and 16.7% of those by the Karhunen-Loeve (K-L) expansion, respectively. Furthermore, EDC can be applied directly to the feature selection of property vectors (PVs) and reduce the measured items of PVs without lowering the identification precision. Compared with the textile yarn grading proposed earlier, EDC provides 16.7% greater efficiency both in measuring PVs and in calculating PAV1 time.

關鍵字(中)

★ 特徵選擇
★ 決策理論
★ 紗線分級

關鍵字(英)

★ Feature Selection
★ Decision Theory
★ Yarn Grading

論文目次

中文摘要 I
英文摘要 Ⅳ
目錄 Ⅶ
表目錄 XII
圖目錄 XIV
縮寫對照 VII
第1章緒論 1
1-1 研究動機 1
1-2 文獻回顧 4
1-3 研究方法及目的 10
1-4 組織架構 12
第2章基本理論 15
2-1 K-L Expansion 15
2-2 Bayes Classifier Method 19
2-3 最小距離法 21
2-4 倒傳遞神經網路 22
第3章以圖形識別理論作紗線分級 26
3-1 前言 26
3-2 實驗數據 28
3-3 K-L Expansion結合Bayes Classifier之學習過程 30
3-4 K-L Expansion結合Bayes Classifier
在學習過程中之特徵選擇 32
3-5 K-L Expansion結合Bayes Classifier
在學習過程中之降階結果 34
3-6 K-L Expansion結合Bayes Classifier之測試過程 37
3-7 K-L Expansion結合Bayes Classifier之測試結果 39
3-8 K-L Expansion結合Minimum Distance作紗線分級 41
3-9 K-L Expansion結合BC或MD作紗線分級之簡易
性與效率性 42
3-10 BC與MD在紗線分級應用上之差異 46
3-11 紗線屬性特徵之指標性 47
3-12 結論 47
第4章以神經網路作紗線分級 50
4-1 前言 50
4-2 實驗流程 52
4-3 倒傳遞神經網路進行紗線等級評估架構 54
4-3-1倒傳遞神經網路學習過程 56
4-3-2倒傳遞神經網路學習結果 58
4-3-3倒傳遞神經網路測試過程 60
4-3-4倒傳遞神經網路測試結果 62
4-4 倒傳遞神經網路與再學習能力評估 63
4-5 K-L Expansion結合倒傳遞神經網路方法進
行紗線等級評估之結果 66
4-5-1 K-L Expansion結合倒傳遞神經網路之特徵選擇 66
4-5-2 K-L Expansion結合倒傳遞神經網路之學習過程 68
4-5-3 K-L Expansion結合倒傳遞神經網路之主軸降階 70
4-5-4 K-L Expansion結合倒傳遞神經網路之測試過程 72
4-5-5 K-L Expansion結合倒傳遞神經網路之測試結果 74
4-6 K-L Expansion結合倒傳遞神經網路進行
紗線等級評估之再學習能力分析 75
4-6-1 K-L Expansion結合神經網路之主軸辨識精度 77
4-6-2 K-L Expansion結合神經網路之再學習能力 78
4-6-3 K-L Expansion結合神經網路之強健性分析 80
4-6-4 K-L Expansion結合神經網路之
再學習收斂速度評估 82
4-7 結論 84
第5章以族群之有效距離作特徵選擇在紗線分級上之應用 87
5-1 前言 87
5-2 以族群之有效距離 (EDC) 作特徵選擇法則 90
5-3 實驗流程 93
5-4 實驗數據 95
5-5 以EDC方法對PAVs作特徵選擇之紗線等級評估 95
5-5-1 PAVs之標準排序與測試排序比較 102
5-5-2 PAVs之誤排距離比較 106
5-5-3 降階過程中之誤判樣本數比較 109
5-5-4 以EDC方法對PAVs作特徵選擇之應用 111
5-6 以EDC方法對PVs作特徵選擇之紗線等級評估 113
5-6-1 PVs之標準排序與測試排序比較 116
5-6-2 以EDC方法對PVs作特徵選擇之誤排距離 119
5-6-3 特徵排序在順向刪除及逆向刪除
相對於誤判樣本數之分析 122
5-6-4 順向及逆向刪除PVs之誤判樣本數比較 124
5-6-5 以EDC方法對PVs作特徵選擇之應用 126
5-7 以EDC方法結合K-L Expansion方法對PVs
作特徵選擇並對PAVs作主軸轉換之紗線等級評估 128
5-7-1 實驗過程 128
5-7-2 順向及逆向刪除PVs之樣本誤判結果 131
5-7-3 EDC方法結合K-L Expansion方法對
PVs作特徵選擇之效益 135
5-7-4 修正以EDC方法結合K-L Expansion方法
作紗線等級評估系統 135
5-7-5 以EDC結合K-L Expansion方法對PVs作
特徵選擇並對PAVs作主軸轉換之應用 137
5-8 結論 141
第6章總結 142
參考文獻 145

參考文獻

1. Annual Book of ASTM Satandards, “Standard Test Method for Grading Spun Yarns for Appearance ASTM D2255-96”, Vol. 07.01,544-548, (1999).
2. Balakrishnan, H., Venkatarman, S., and Jayaraman, S., FDICS: A Vision-based System for the Identification and Classification of Fabric, J. Text. Inst. 89, (2), 365-380 (1998).
3. Basak, J., and Mitra, S., Feature Selection Using Radial Basis Function Networks, Neural Computing & Applications, 8, 297-302 (1999).
4. Basak J, De RK, and Pal SK., Fuzzy feature evaluation index and connbectionist realization –II: Theoretcal analysis, Information Sciences, 111,1-17 (1998).
5. Basak J, De Rk, Pal SK., Unsupervised feature selection using neurofuzzy approach, Pattern Recognition Letters, 19, 997-1006(1998).
6. Bezdek J.C., and Castelaz, P.F., Prototype Classification and Feature Selection with Fuzzy Sets, IEEE Trans Sys. Man and Cybernetics, 7, 87-92 (1997).
7. Chen, Y., and Collier, B. J., Characterizing Fabric End-Use by Fabric Physical Properties, Textile Res. J. 67, (4), 247-252 (1997).
8. Chen, Y., Collier, B., Hu, P., and Quebedeaux, D., Objective Evaluation of Fabric Softness, Textile Res. J. 70(5), 443-448 (2000).
9. Cheng, L., and Adams, D.L., Yarn Strength Prediction Using Neural Networks Part I: Fiber Properties and Yarn Strength Relationship, Textile Res. J. 65(9), 495-500 (1995).
10. Devijer, P.A. and Kittler, J. “Pattern Recognition: A Statistical Approach, ” Prentice-Hall, NY, 1982.
11. Fan, K. C., and Wang, Y. K., Fabric Classification Based on Recognition Using a Neural Network and Dimensionality Reduction, Textile Res. J. 68, (3), 179-185(1998).
12. Hunter, A., Feature Selection Using Probabilistic Neural Networks, Neural Computing & Applications, 9, 124-132 (2000).
13. Kawabata, S., The Standardization and Analysis of Hand Evaluation, HESC, the Textile Machinery Society of Japan, Osaka, Japan, 2nd edition, 1980.
14. Kawabata. S., Ed., “HESC Standard of Hand Evaluation （H. V. Standard for Men’s Suiting）, ” HESC, The Textile Machinery Society of Japan, Osaka, Japan, 1975.
15. Kawabata. S., Ed., “HESC Standard of Hand Evaluation. ” vol. 1 （H.V. Standard for Men’s Suiting）. 2nd ed., HESC. The Textile Machinery Society of Japan, Osaka, Japan, 1980.
16. Kawabata, S., Ed., ”HESC Standard of Hand Evaluation, ” vol.2（H. V. Standard for Women’s Thin Dress Fabric）, HESC, The Textile Machinery Society of Japan, Osaka, Japan, 1980.
17. Kawabata, Sueo, “The Standardization and Analysis of Hand Evaluation, ” 2nd ed., The Hand Evaluation and Standardization Committee. The Textile Machinery Society of Japan, 1980.
18. Kawabata, S., Ed “HESC Standard of Hand Evaluation, ” vol.3（T. H. V. Standard for Men’s Winter Suiting）, HESC, The Textile Machinery Society of Japan, Osaka, Japan, 1982.
19. Kawabata. S., “Development of the Objective Measurement of Fabric Handle, Objective Specification of Fabric Quality, Mechanical Properties and Performance, ” S. Kawabata. R. Postle, and Masako Niwa, Eds. Proc. 1st Japan –Australia Joint Symposium, The Textile Machinery Society of Japan, 31-59,1982.
20. Kawabata, S., Development of a Fabric Propetry Measurement, Part I: A New Compression Tester, report of a research project (grant-in-aid for co-operative research (A), Application of the Objective Measurement of Fabric Propetry to the Design of Fabric Quality’, head investigator Masako Niwa, project no. 6130012, supported by the Japanese Scientific Fund), 156-161,1988.
21. Kawabata, S., Development of a Fabric Property Measurement, Part II: A Surface Analyzer, report of a research project (grant-in-aid for co-operative research (A), Application of the Objective Measurement of Fabric Property to the Design of Fabric Quality, head investigator Masako Niwa, project no. 6130012, supported by the Japanese Scientific Fund), 162-165, 1988
22. Kawabata, S., and Niwa, M., Fabric Performance in Clothing and Clothing Manufacture, J. Textile Inst. 80(1), 19-50 (1989).
23. Kawabata, S., and Niwa, Masako, The Measurement of Fabric Characteristics and Surface Properties, in “Proc.index ’93 Congress.”5D. The Research and Development Session, Research on Nonwoven Materials, 1993.
24. Kawabata, S., Niwa, M., and Wang, F., Objective Hand Measurement of Nonwoven Fabrics, Part I: Developmentn of the Equations, Textile Res. J. 64(10), 597-610(1994).
25. Laurene F., Fundamentals of Neural Networks. Englewood Cliffs, NJ: Prentice-Hall, 1994.
26. Lien H.C., and Lee S., Application of Pattern Recognition Principles for Grading Textile Yarns, Textile Res. J. 72(4), 320-326 (2002).
27. Lien, H.C., and Lee, S., A Method of Feature Selection for Textile Yarn Grading Using the Effective Distance Between Clusters, Textile Res. J. 72(10), 870-878 (2002).
28. Ludwig, L., Sapozhnikova, E., Lunin, V., and Rosenstiel, W., Error Classification and Yield Prediction of Chips in Semiconductor Industry Applications, Neural Computing & Applications, 9, 202-210 (2000).
29. Lawson-Hemphill Inc., The Electronic Inspection Board Operator’s Manual, Doc 481-100, USA. , 1995.
30. Lin, J. J., Lin, C. H., and Tsai, I. S., Applying Expert System and Fuzzy Logic to an Intelligent Diagnosis System for Fabric Inspection, Textile Res. J. 65, (12), 697-709 (1995).
31. Mao, J., and Jain, A.K., Artificial Neural Networks for Feature Extraction and Multivariate Data Projection, IEEE Trans Neural Networks, 6, 296-317 (1995).
32. Niwa, M., Liu, Caiming, and Kawabata, S., Application of the Objective Fabric Hand Evaluation Technology to Other Materials Such as Artificail Leather, Foam, and Tissue Paper, in “Proc. 18th Textile Tech. Symposium, ” The Textile Machinery Soc. Japan, 167-173 (1989).
33. Niwa, M., Inoue, M., and Kawabata, S., The Objective Evaluation of Blanket Hand and Durability, A Preliminary Investigation, Int. J. Cloth. Sci Technol. 11, 90-104(1999).
34. Pal SK, Chakraborty B., Fuzzy set theoretic measure for automatic feature evaluation, IEEE Trans Systems, Man and Cybernetics, 16, 754-760 (1986).
35. Pal SK., Fuzzy set theoretic measures for automatic feature evaluation: II, Information Sciences, 64, 165-179 (1992).
36. Pal SK., Basak j, De RK, Fuzzy feature evaluation index and connectionist realization, Information Sciences, 105, 173-188 (1998).
37. Park, C.K., and Kang, T.J., Objective Rating of Seam Pucker Using Neural Networks, Textile Res. J. 67(7), 494-502 (1997).
38. Park, W.S., Hwang, Y.G., Kang, B.C., and Yeo, S.W., Applying Fuzzy Logic and Neural Networks to Total Hand Evaluation of Knitted Fabrics, Textile Res. J. 70(8). 675-681 (2000).
39. Pan, N., Yen, K.C., Zhao, S.J., and Yang, S.R., A New Approach to the Objective Evaluation of Fabric Handle from Mechanical Properties Part II：Objective Measures for Primary Handle, Textile Res. J., 531-537 (1988).
40. Pynckels, F., Kiekens, P., Sette, S., Van Langgenhove, L., and Impe, K., The use of Neural Nets to Simulate the Spinning Process, J. Text. Inst. 88, (4), 440-448 (1997).
41. Priddy KL, Rogers SK, Ruck DW, Tarr GL, Kabrisky M., Bayesian selection of important features for feedforward neural networks. Neurocomputing, 5, 91-103 (1993).
42. Raheel, M., and Liu, Jiang, An Empirical Model for Fabric Hand Part I：Objective Assessment of Light Weight Fabrics, Textile Res. J. 61(1), 31-38 (1991).
43. Raheel, M., and Liu, J., An Empirical Model for Fabric Hand Part II：Subjective Assessment, Textile Res. J. 61(2), 79-82 (1991).
44. Rong, G.H., Slater, K.L., and Fei, R.C., The Use of Cluster Analysis for Grading Textile Yarns, J. Textile Inst. 85(3), 389-396 (1994).
45. Radhakrishnaiah, P., Tejatanalert, and Sawhney, A.P.S., Handle and Comfort Properties of Woven Fabrics Made from Random Blend and Coton-Covered Cotton/Polyester Yarns, Textile Res. J. 63(10), 573-579 (1993).
46. Ruck Dw, Rogers SK, and Kabrisky M., Feature selection using a multilayer perceptron, Neural Network Computing, 20, 40-48 (1990).
47. Sette, S., Boullart, L., Van Langenhov, L., and Kiekens, P., Optimizing the Fiber-to-Yarn Production Process with a Combined Neural Network/Genetic Algorithm Approach, Textile Res. J. 67(2), 84-92 (1997).
48. Sick B., On-Line Tool Wear Monitoring Using Neural Networks. Neural Computing & Applications, 7, 356-366 (1998).
49. State Stand Bureau, P. R. China Standard：Combed Polyester and Cotton Blended Yarns, 1985.
50. Tou, J.T., and Heydorn, P.R., Some Approaches to Optimum Feature Extraction, in “Computer and Information Sciences-II,” Academic Press, N.Y., 1967.
51. Tou, J.T., and Gonzalez, R.C., “ Pattern Recognition Principles, ” Addison-Wesley, Reading, MA, 1974.
52. Verikas, A., Malmqvist, K., Bergman, L., and Signahl, M., Colour Classification by Neural Networks in Graphic Arts, Neural Computing & Applications, 7, 52-64 (1998).

指導教授

李雄(Shyong Lee)

審核日期

2003-7-1

推文