single-jc.php

JACIII Vol.11 No.6 pp. 620-625
doi: 10.20965/jaciii.2007.p0620
(2007)

Paper:

Neighbor Size of Initial Node Exchange and its Influence for SOM Learning

Tsutomu Miyoshi

Department of Information and Knowledge Engineering, Tottori University, 4-101 Koyama-cho Minami, Tottori-shi, Tottori 680-8552, Japan

Received:
January 12, 2007
Accepted:
March 20, 2007
Published:
July 20, 2007
Keywords:
self organizing map, learning method, improvement, node exchange
Abstract

The Self Organizing Map (SOM) involves neural networks, that learns the features of input data through unsupervised, competitive neighborhood learning. In the SOM learning algorithm, connection weights in a SOM feature map are initialized at random values, which also sets nodes at random locations in the feature map independent of input data space. The distance that output nodes move increases, slowing learning convergence. We propose solving this problem in initial node exchange using part of learning data. We investigated how the average move distance of all nodes, measured by convergence, changed with differences in the initial size of the neighbor area in node exchange. We clarified the influence of the initial neighbor size on the average move distance of all nodes, clarified the expression of relations, and showed the optimum domain of relations.

Cite this article as:
Tsutomu Miyoshi, “Neighbor Size of Initial Node Exchange and its Influence for SOM Learning,” J. Adv. Comput. Intell. Intell. Inform., Vol.11, No.6, pp. 620-625, 2007.
Data files:
References
  1. [1] T. Kohonen, “Self-Organizing Maps,” Springer Verlag,
    ISBN 3540586008, 1995.
  2. [2] R. Heclt-Nielsen, “Neurocomputing,” Addison-Wesley Pub. Co.,
    ISBN 0-201-09355-3, 1990.
  3. [3] “Bibliography of SOM papers,” Draft version 2006-03-27,
    http://www.cis.hut.fi/mpolla/sombibliography/
  4. [4] M. Kinouchi and Y. Kudo, “Much faster learning algorithm for batch-learning som and its application to bioinformatics,” Proceedings of The Workshop on Self-Organizing Maps (WSOM03), 2003.
  5. [5] Y. P. Jun, H. Yoon, and J. W. Cho, “L* learning: a fast selforganizing feature map learning algorithm based on incremental ordering,” IEICE Transactions on Information and Systems, Vol.E76-D, No.6, pp. 698-706, 1993.
  6. [6] H. Tamukou, K. Horio, and T. Yamakawa, “Fast learning algorithms for self-organizing map employing rough comparison WTA and its digital hardware implementation,” IEICE Transactions on Electronics, Vol.E87-C, No.11, pp. 1787-1794, 2004.
  7. [7] T. Miyoshi, H. Kawai, and H. Masuyama, “Efficient SOM Learning by Data Order Adjustment,” Proceedings of 2002 IEEE World Congress on Computational Intelligence (WCCI2002), p. 784, 2002.
  8. [8] T. Miyoshi, “Order of Learning Data and Convergence of SOM Learning,” Proceedings of The 6th International Symposium on Advanced Intelligent Systems (ISIS2005), pp. 756-759, 2005.
  9. [9] T. Miyoshi, “Learning Data Order and Convergence of SOM Learning,” GESTS International Transactions on Computer Science and Engineering, Vol.22, No.1, pp. 188-197, 2005.
  10. [10] T. Miyoshi, “Node Exchange for Improvement of SOM Learning,” Proceedings of 9th International Conference on Knowledge-Based Intelligent Information and Engineering Systems (KES2005), pp. 569-574, 2005.
  11. [11] T. Miyoshi, “Initial Node Exchange and Convergence of SOM Learning,” Proceedings of The 6th International Symposium on Advanced Intelligent Systems (ISIS2005), pp. 316-319, 2005.
  12. [12] T. Miyoshi, “Initial Node Exchange Using Learning Data and Convergence of SOM Learning,” GESTS International Transactions on Computer Science and Engineering, Vol.21, No.1, pp. 216-224, 2005.

*This site is desgined based on HTML5 and CSS3 for modern browsers, e.g. Chrome, Firefox, Safari, Edge, Opera.

Last updated on May. 10, 2021