臺灣博碩士論文加值系統

混合型的資料集─混合數值型及種類型屬性的資料集，在現代的企業資料庫中是很常見的；其中珍貴的特徵可能是不容易被察覺的。而分析混合型的數據來抽取這些不容易察覺的資料特徵，對於決策制訂的過程是非常可貴的；而這重要的關鍵因素更能幫助組織維持其競爭力;此外，視覺化分析更能提升初始資料分析期間的分析品質。在本論文中，我們提出了一個基於應用擴充式的自組映射圖，能分析多屬性混合型態資料的視覺化分析架構；這個架構能將資料分群分析及分類分析，視覺化地表現在投射後的地圖上。我們通過分析兩個真實世界的資料集、並比較其他現有的模式以顯示其優勢，從而驗證了該方法的可行性。

Mixed numeric and categorical data are commonly seen in nowadays corporate databases in which precious patterns may be hidden. Analyzing mixed-type data to extract the hidden patterns valuable to decision-making is therefore beneficial and critical for corporations to remain competitive. In addition, visualization facilitates exploration in the early stage of data analysis. In the paper, we present a visualized approach for analyzing multivariate mixed-type data. The proposed framework based on an extended self-organizing map allows visualized data cluster analysis as well as classification. We demonstrate the feasibility of the approach by analyzing two real-world datasets and compare our approach with other existing models to show its advantages.

Table of Contents
摘要 ii
Abstract iii
Table of Contents iv
List of Figures vi
List of Tables vii
1. Introduction 1
1.1. Motivation 1
1.2. Objective 2
1.3. Organization 2
2. Literature Review 3
2.1. Background 3
2.1.1. SOM 3
2.2. Related Work 4
2.2.1. GSOM 4
2.2.2. Visualization-Induced SOM 4
2.2.3. Problems of the conventional SOM and ViSOM 5
2. Methodology 7
3.1 The Extended SOM 7
3.1.1. Distance hierarchy. 7
3.1.2. Mapping an input to distance hierarchies. 9
3.1.3. Mapping a neuron prototype to distance hierarchies. 9
3.1.4. Aggregate the distances. 10
3.2 Decomposing updating force on distance hierarchy 11
3.3 Adjusting a MixSOM neuron 13
3.4 Construct Distance Hierarchies 16
3.5 Classification by Using Trained MixSOM 17
4. Experimental results 20
4.1 Synthetic dataset - NomOrdNum 20
4.2 Real dataset - Adult 23
4.3 Real dataset - Australian Credit Approval 29
4.4 Classification Performance 32
4.4.1 Adult Dataset 32
4.4.2 Credit Approval Dataset 34
5. Conclusion 36
References 37


List of Figures

Fig. 1. Distance hierarchies for nominal, ordinal, and numeric attributes. 7
Fig. 2. Decomposition of updating forces when mapping points X, K, and V are on the same path (a-c) and on different paths (d-f). 12
Fig. 3. The adjustment of a prototype-element mapping point M toward a data-element mapping point X. 14
Fig. 4. Projection results of the synthetic dataset NomOrdNum 21
Fig. 5. Automatically constructed distance hierarchies for the categorical attributes of dataset Adult. The max distance between two values in each attribute is normalized to 1. 24
Fig. 6. Projection results of the Adult training dataset 25
Fig. 7. Clustering with class distribution reflected by the grey-level of neurons for the Adult training dataset 26
Fig. 8. Automatically constructed distance hierarchies for the categorical attributes of dataset Credit Approval. 30
Fig. 9. Projection results of the training data of Credit Approval 31
Fig. 10. Classification accuracy by the instance-based approach using k nearest neighbors in GSOM/MixSOM neurons or in the training dataset (indicated by the kNN) 33


List of Tables

Table 1. The corresponding matrix of the hierarchy in Fig. 3a 15
Table 2. The synthetic mixed-type dataset NomOrdNum 20
Table 3. Gain ratio on class Salary after data partitioning according to the trained maps 28
Table 4. Characteristics of each cluster in Fig. 7(b) sorted by % of salary class >50K 28
Table 5. Characteristics of each cluster in Fig. 9(c) sorted by % of class label 32
Table 6. Gain ratio on class Attribute after data partitioning according to the trained maps 32
Table 7. Classification accuracy by the neuron-based approach with which the neurons within distance d to the bmu participating in the voting 32
Table 8. The Best accuracy by the neuron-based and the instance-based approach and their parameter settings on dataset adult 33
Table 9. Classification accuracy by the neuron-based approach with which the neurons within distance d to the bmu participating in the voting. 34
Table 10. The Best accuracy by the neuron-based and the instance-based approach and their parameter settings on dataset Credit approval 34
Table 11. 10-fold cross-validation results of dataset Credit Approval 35

References
[1]K. Pearson, "On lines and planes of closest fit to systems of points in space," Philosophical Magazine Series 6, vol. 2, pp. 559-572, 1901.
[2]J. W. Sammon Jr, "A nonlinear mapping for data structure analysis," IEEE Transactions on computers, vol. 100, pp. 401-409, 1969.
[3]I. Borg and P. J. F. Groenen, Modern multidimensional scaling: Theory and applications, (2nd ed.) ed. New York: Springer Verlag, 2005.
[4]C. C. Hsu, K. M. Wang, and S. H. Wang, "GViSOM for Multivariate Mixed Data Projection and Structure Visualization," in IJCNN, 2006, pp. 3300-3305.
[5]T. Kohonen, "The self-organizing map," Proceedings of the IEEE, vol. 78, pp. 1464-1480, 1990.
[6]C. C. Hsu, "Generalizing self-organizing map for categorical data," IEEE transactions on Neural Networks, vol. 17, pp. 294-304, 2006.
[7]H. Yin, "Data visualisation and manifold mapping using the ViSOM," Neural Networks, vol. 15, pp. 1005-1016, 2002.
[8]R. Mayer, T. Abdel Aziz, and A. Rauber, "Visualising class distribution on self-organising maps," Lecture Notes in Computer Science, vol. 4669, pp. 359-368, 2007.
[9]K. Tasdemir and E. Merenyi, "Exploiting data topology in visualization and clustering of Self-Organizing Maps," IEEE Transactions on Neural Networks, vol. 20, p. 549, 2009.
[10]H. Yin, "ViSOM-a novel method for multivariate data projection and structure visualization," IEEE transactions on Neural Networks, vol. 13, pp. 237-243, 2002.
[11]J. Vesanto and E. Alhoniemi, "Clustering of the self-organizing map," IEEE transactions on Neural Networks, vol. 11, pp. 586-600, 2000.
[12]M. Aupetit, "Visualizing the trustworthiness of a projection," in Proc. 14th European Symposium of Artificial Neural Networks (ESANN06), 2006, pp. 271-276.
[13]M. Cottrell and E. de Bodt, "A Kohonen map representation to avoid misleading interpretations," in Proc. European Symposium on Artificial Neural Networks, Bruges, Belgium, D-Facto, 1996, pp. 103-110.
[14]D. Merkl and A. Rauber, "Alternative ways for cluster visualization in self-organizing maps," in Proc. Workshop on Self-Organizing Maps (WSOM97), Espoo, Finland, 1997, pp. 106-111.
[15]G. Polzlbauer, A. Rauber, and M. Dittenbach, "Advanced visualization techniques for self-organizing maps with graph-based methods," in In Proc. Int. Symp. On Neural Networks (ISSN05), 2005, pp. 75-80.
[16]K. Tasdemir, "Exploring Topology Preservation of SOMs with a Graph Based Visualization," in Proc. Int. Conf. on Intelligent Data Engineering and Automated Learning, 2008, pp. 180-187.
[17]A. Ultsch, "Self-organizing neural networks for visualization and classification," Information and classification, pp. 864-867, 1993.
[18]H. Ressom, D. Wang, and P. Natarajan, "Adaptive double self-organizing maps for clustering gene expression profiles," Neural Networks, vol. 16, pp. 633-640, 2003.
[19]S. Wu and T. W. S. Chow, "Clustering of the self-organizing map using a clustering validity index based on inter-cluster and intra-cluster density," Pattern Recognition, vol. 37, pp. 175-188, 2004.
[20]J. Lampinen and E. Oja, "Clustering properties of hierarchical self-organizing maps," Journal of Mathematical Imaging and Vision, vol. 2, pp. 261-272, 1992.
[21]M. Y. Kiang, "Extending the Kohonen self-organizing map networks for clustering analysis," Computational Statistics and Data Analysis, vol. 38, pp. 161-180, 2001.
[22]F. Murtagh, "Interpreting the Kohonen self-organizing feature map using contiguity-constrained clustering," Pattern Recognition Letters, vol. 16, pp. 399-408, 1995.
[23]G. Das, H. Mannila, and P. Ronkainen, "Similarity of attributes by external probes," Knowledge Discovery and Data Mining, pp. 23–29, 1998.
[24]C. R. Palmer and C. Faloutsos, "Electricity based external similarity of categorical attributes," Lecture notes in computer science, pp. 486-500, 2003.
[25]T. Kohonen, J. Kangas, and J. Laaksonen, "SOM_PAK: The Self-Organizing Map Program Package, Ver 3.1," SOM programming team of Helsinki University of Technology, 1995.
[26]L. T. Ho, "Clustering of Generalized Visualization-induced SOM on Mixed Data," in Institute of Information Management. Unpublished master''s thesis Douliu, Yunlin, Taiwan.: National Yunlin University of Science & Technology, 2008, p. 40.
[27]C. L. Blake and C. J. Merz, "UCI repository of machine learning databases. Univ. California, Dept," Inform. Comput. Sci., Irvine, CA.[Online]. Available: http://www.ics.uci.edu/~mlearn/ML-Repository.html, 1998.
[28]J. R. Quinlan, C4. 5: programs for machine learning: Morgan Kaufmann, 2003.
[29]Y. Peng, G. Kou, Y. Shi, and Z. Chen, "A Multi-criteria Convex Quadratic Programming model for credit data analysis," Decision Support Systems, vol. 44, pp. 1016-1030, 2008.