臺灣博碩士論文加值系統

自組織映射圖（Self-organizing map，SOM）是一種人工神經網路（ANN）的類型， 是一種非監督學習，以離散的表示訓練樣本的輸入空間，稱為地圖，並應用競爭性學習並使用鄰域函數來保留輸入空間的拓撲性質。
由於SOM會產生拓撲空間，其搜索與學習的性質需要大量的學習次數來進行收斂，資料輸入順序可能影響SOM之單元學習，若某一個單元贏得連續數筆相近似之輸入資料(input data)，該單元可能快速朝該幾筆資料靠攏，結果可能使SOM之學習快速收斂，可能陷入局部最佳化，但亦可能逃脫局部最佳化。
本研究目的在了解輸入順序對於自組織映射圖學習的影響。我們採用主成分分析(Principal components analysis，PCA)和多維尺度(Multidimensional scaling，MDS )於決定輸入資料之順序。以主成分分析中，資料會依主成分進行投影，本研究取原始資料經由PCA投影後之第一主成份順序進入自組織映射圖。在利用多維尺度時，將資料投影至一維空間，依資料在此一維空間之順序，輸入自組織映射圖。
本研究採用的評量方式為量化誤差、群間和群內距離作為比較的依據。量化誤差是以計算並加總輸入向量和優勝神經元的權重向量的歐基里德距離，群間距離差異大資料結果越佳，群內距離差異小資料結果越佳之特性，利用此三個數據獲得資料輸入順序對自組織映射圖的學習之影響。資料經過主成分分析與多維尺度後之順序再輸入自組織映射圖，與原始的資料順序比較後，資料經過主成分分析與多維尺度後之順序再輸入自組織映射圖，在相同的學習速率下，會有比較穩定的收斂情況，其迭代所需要的回合數也較原始資料順序少。

中文摘要 i
Abstract iii
誌謝 v
目錄 vi
圖目錄 viii
表目錄 ix
第一章、 緒論 1
1.1 研究背景與動機 1
1.2 論文架構 2
第二章、 相關研究 3
2.1 主成分分析 3
2.2 多維尺度 5
2.3 自組織映射圖 5
第三章、 研究架構 8
3.1 主成分分析(Principal components analysis, PCA) 8
3.2 多維尺度(Multidimensional scaling) 12
3.3 自組織映射圖 15
第四章、 實驗結果與分析 19
4.1 評估方法 19
4.2 資料集 20
4.3 Wine 20
4.4 Mushroom 21
4.5 實驗結果 22
第五章、 結論與未來 26
5.1 結論 26
5.2 未來展望 27
參考文獻 28

[1] H. Abdi, "Partial least squares regression and projection on latent structure regression (PLS Regression), " Wiley Interdisciplinary Reviews: Computational Statistics, pp. 97-106, 2010.

[2] H. Abdi & L. J. Williams, "Principal component analysis, " Wiley interdisciplinary reviews: computational statistics, pp. 433-459, 2010.

[3] E. Achtert, C. Böhm & P. Kröger, "DeLi-Clu: boosting robustness, completeness, usability, and efficiency of hierarchical clustering by a closest pair ranking, " Pacific-Asia Conference on Knowledge Discovery and Data Mining, pp.119-128, 2006.

[4] A. A. Akinduko, E. M. Mirkes & A. N. Gorban, "SOM: Stochastic initialization versus principal components, " Information Sciences, pp. 213-221, 2016.

[5] T. P. Barnett, & R. Preisendorfer, "Origins and levels of monthly and seasonal forecast skill for United States surface air temperatures determined by canonical correlation analysis, " Monthly Weather Review, pp. 1825-1850, 1987.

[6] I. Borg & P. J. Groenen, "Modern multidimensional scaling: Theory and applications, " Springer Science & Business Media, 2005.

[7] A. M. Bronstein, M. M. Bronstein, & R. Kimmel, "Generalized multidimensional scaling: a framework for isometry-invariant partial surface matching, " Proceedings of the National Academy of Sciences, 2006.

[8] A. P. M. Coxon & P. M. Davies, The user's guide to multidimensional scaling: With special reference to the MDS (X) library of computer programs, Heinemann, 1982.

[9] G. Deco & D. Obradovic, An information-theoretic approach to neural computing, Springer Science & Business Media, 2012.

[10] L. R. Fabrigar, D. T. Wegener, R. C. MacCallum & E. J. Strahan, "Evaluating the use of exploratory factor analysis in psychological research, " Psychological methods, 1999.

[11] A. Gersho & R. M. Gray, "Vector quantization and signal compression, " Springer Science & Business Media, Vol. 159, 2012.

[12] D. L. Hernández-Rabadán, J. Guerrero, & F. Ramos-Quintana, "Method for segmenting tomato plants in uncontrolled environments, " Engineering, Vol. 4, pp. 599-606, 2012.

[13] G. E. Hinton & T. J. Sejnowski, Unsupervised learning: foundations of neural computation, MIT Press, 1999.

[14] I. T. Jolliffe, "Principal components as a small number of interpretable variables: some examples, " Principal Component Analysis, pp. 63-77, 2002

[15] T. Kohonen, & T. Honkela, "Kohonen network, " Scholarpedia, 2007.


[16] S. B. Kotsiantis, I. Zaharakis & P. Pintelas, "Supervised machine learning: A review of classification techniques, " Emerging artificial intelligence applications in computer engineering, 2007.

[17] J. B. Kruskal, "Multidimensional scaling by optimizing goodness of fit to a nonmetric hypothesis, " Psychometrika, 1964.

[18] K. LIII. Pearson, "On lines and planes of closest fit to systems of points in space, " The London, Edinburgh, and Dublin Philosophical Magazine and Journal of Science, pp. 559-572, 1901.

[19] S. T. Roweis, "EM algorithms for PCA and SPCA, " Advances in neural information processing systems, pp. 626-632, 1998.

[20] A. A. Sodemann, M. P. Ross, & B. J. Borghetti, "A review of anomaly detection in automated surveillance, " IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), 2012.

[21] A. Ultsch, "U*-matrix: a tool to visualize clusters in high dimensional data, " Marburg: Fachbereich Mathematik und Informatik, 2003.

[22] F. Wickelmaier, "An introduction to MDS, " Sound Quality Research Unit, Aalborg University, Denmark, 2003.

[23] H. Yin, "Learning nonlinear principal manifolds by self-organising maps, " Principal manifolds for data visualization and dimension reduction, pp. 68-95, 2008.