Abstract
Kohonen’s Self-organizing Map (SOM) is one of the most popular neural network algorithms. SOM produces topology preserving map of the input data. In the current study the SOM’s topology preservation property is used to identify the input features whose removal does not affect significantly the neighborhood relations among the input data points. The topology preservation property of of an SOM is measured using a quantitative index. However the same index can be slightly modified to compute topology preservation in the SOM along individual features. Thus studying the topology preservation due to each individual feature we can compare their quality with respect to their importance in affecting the neighborhood relation among input points. Experimental study is conducted with a synthetic data set, well known Iris data set and a multi-channel satellite image dataset. The results are cross verified by comparing with Sammon error of the data computed in the corresponding dimension. k-NN classification performance is also considered for the data sets.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Similar content being viewed by others
References
Kohonen, T.: Self-Organizing Maps. Springer Series in Information Sciences, vol. 30. Springer, Berlin (1995)
Kohonen, T., Oja, E., Simula, O., Visa, A., Kangas, J.: Engineering application of self-organizing map. Proc. IEEE 84(10), 1358–1383 (1996)
Ritter, H., Schulten, K.: Kohonen’s self-organizing maps: exploring their computational capabilities. In: IEEE Int. Conf. on Neural Networks, San Diego, vol. 1, pp. 109–116 (1988)
Himberg, J., Ahola, J., Alhoniemi, E., Vesanto, J., Simula, O.: The self-organizing map as a tool in knowledge engineering. In: Pal, N.R. (ed.) Pattern Recognition in Soft Computing Paradigm, pp. 38–65. World Scientific, Singapore (2001)
Bauer, H., Pawelzik, K.R.: Quantifying the Neighborhood Preservation of Self-Organizing Feature Maps. IEEE Trans. on Neural Networks 3(4), 570–579 (1992)
Villmann, T., Der, R., Herrmann, M., Martinetz, T.M.: Topology preservation in self-organizing feature maps: exact definition and measurements. IEEE Trans. on Neural Networks 8(2), 256–266 (1997)
Laha, A., Pal, N.R.: On different variants of self-Organizing feature map and their properties. In: Proceedings of the 1999 IEEE Hong Kong Symp. on Robotics and Controls, vol. 1, pp. I-344–I-349 (1999)
Su, M.C., Chang, H.T., Chou, C.H.: A novel measure for quantifying the topology preservation of self-organizing feature maps. Neural Processing Letters 15(2), 137–145 (2002)
Fukunaga, K.: Introduction to Statistical Pattern Recognition. Academic Press Inc., Harcourt Brace Jovanovich Publishers (1990)
Ritter, H., Schulten, K.: Convergence properties of Kohonen’s topology conserving maps: Fluctuations, stability and dimension selection. Biol. Cybern. 60, 59–71 (1988)
Kwak, N., Choi, C.H.: Input feature selection for classification problems. IEEE Trans. on Neural Networks 13(1), 143–159 (2002)
Sammon, J.W.: A nonlinear mapping for data structure analysis. IEEE Trans. Comput. 18, 401–409 (1969)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2004 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Laha, A. (2004). Detecting Topology Preserving Feature Subset with SOM. In: Das, G., Gulati, V.P. (eds) Intelligent Information Technology. CIT 2004. Lecture Notes in Computer Science, vol 3356. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-30561-3_5
Download citation
DOI: https://doi.org/10.1007/978-3-540-30561-3_5
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-24126-3
Online ISBN: 978-3-540-30561-3
eBook Packages: Computer ScienceComputer Science (R0)