Advertisement

Fast BMU Search in SOMs Using Random Hyperplane Trees

  • César A. Astudillo
  • B. John Oommen
Part of the Lecture Notes in Computer Science book series (LNCS, volume 8862)

Abstract

One of the most prominent Neural Networks (NNs) reported in the literature is the Kohonen’s Self-Organizing Map (SOM). In spite of all its desirable capabilities and the scores of reported applications, it, unfortunately, possesses some fundamental drawbacks. Two of these handicaps are the quality of the map learned and the time required to train it. The most demanding phase of the algorithm involves determining the so-called Best Matching Unit (BMU), which requires time that is proportional to the number of neurons in the NN. The focus of this paper is to reduce the time needed for this tedious task, and to attempt to obtain an approximation of the BMU is as little as logarithmic time. To achieve this, we depend heavily on the work of [3,6], where the authors focused on how to accurately learn the data distribution connecting the neurons on a self-organizing tree, and how the learning algorithm, called the Tree-based Topology-Oriented SOM (TTOSOM), can be useful for data clustering [3,6] and classification [5]. We briefly state how we intend to reduce the training time for identifying the BMU efficiently. First, we show how a novel hyperplane-based partitioning scheme can be used to accelerate the task. Unlike the existing hyperplane-based partitioning methods reported in the literature, our algorithm can avoid ill-conditioned scenarios. It is also capable of considering data points that are dynamic. We demonstrate how these hyperplanes can be recursively defined, represented and computed, so as to recursively divide the hyper-space into two halves. As far as we know, the use of random hyperplanes to identify the BMU is both pioneering and novel.

Keywords

Approximate BMU search Random Hyperplane Trees SOM TTOSOM VQ 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Astudillo, C.A., Oommen, B.J.: A novel self organizing map which utilizes imposed tree-based topologies. In: 6th International Conference on Computer Recognition Systems, vol. 57, pp. 169–178 (2009)Google Scholar
  2. 2.
    Astudillo, C.A., Oommen, B.J.: On using adaptive binary search trees to enhance self organizing maps. In: Nicholson, A., Li, X. (eds.) AI 2009. LNCS, vol. 5866, pp. 199–209. Springer, Heidelberg (2009)CrossRefGoogle Scholar
  3. 3.
    Astudillo, C.A., Oommen, B.J.: Imposing tree-based topologies onto self organizing maps. Information Sciences 181(18), 3798–3815 (2011)CrossRefMathSciNetGoogle Scholar
  4. 4.
    Astudillo, C.A., Oommen, B.J.: Semi-supervised classification using tree-based self-organizing maps. In: Wang, D., Reynolds, M. (eds.) AI 2011. LNCS, vol. 7106, pp. 21–30. Springer, Heidelberg (2011)CrossRefGoogle Scholar
  5. 5.
    Astudillo, C.A., Oommen, B.J.: On achieving semi-supervised pattern recognition by utilizing tree-based soms. Pattern Recognition 46(1), 293–304 (2013)CrossRefzbMATHGoogle Scholar
  6. 6.
    Astudillo, C.A., Oommen, B.J.: Self-organizing maps whose topologies can be learned with adaptive binary search trees using conditional rotations. Pattern Recognition 47(1), 96–113 (2014)CrossRefGoogle Scholar
  7. 7.
    Astudillo, C.A., Oommen, B.J.: Topology-oriented self-organizing maps: A survey. Pattern Analysis and Applications (2014), http://dx.doi.org/10.1007/s10044-014-0367-9
  8. 8.
    Bentley, J.L.: Multidimensional binary search trees used for associative searching. Commun. ACM 18(9), 509–517 (1975)CrossRefzbMATHMathSciNetGoogle Scholar
  9. 9.
    Chen, S.X., Li, F.W., Zhu, W.L.: Fast searching algorithm for vector quantisation based on features of vector and subvector. IET Image Processing 2(6), 275–285 (2008)CrossRefMathSciNetGoogle Scholar
  10. 10.
    Devroye, L., King, J., McDiarmid, C.: Random hyperplane search trees. SIAM J. Comput. 38(6), 2411–2425 (2009)CrossRefzbMATHMathSciNetGoogle Scholar
  11. 11.
    Frank, A., Asuncion, A.: UCI machine learning repository (2010), http://archive.ics.uci.edu/ml
  12. 12.
    Friedman, M., Kandel, A.: Introduction to Pattern Recognition: Statistical, Structural, Neural and Fuzzy Logic Approaches. Imperical College Press (1999)Google Scholar
  13. 13.
    Gordon, L., Olshen, R.A.: Asymptotically efficient solutions to the classification problem. Annals of Statistics 6, 515–533 (1978)CrossRefzbMATHMathSciNetGoogle Scholar
  14. 14.
    Gray, R.: Vector quantization. IEEE ASSP Magazine 1(2), 4–29 (1984)CrossRefGoogle Scholar
  15. 15.
    Kasai, W., Tobe, Y., Hasegawa, O.: A fast BMU search for support vector machine. In: Alippi, C., Polycarpou, M., Panayiotou, C., Ellinas, G. (eds.) ICANN 2009, Part I. LNCS, vol. 5768, pp. 864–873. Springer, Heidelberg (2009)CrossRefGoogle Scholar
  16. 16.
    Koikkalainen, P., Oja, E.: Self-organizing hierarchical feature maps. In: IJCNN International Joint Conference on Neural Networks, vol. 2, pp. 279–284 (June 1990)Google Scholar
  17. 17.
    Lai, J.Z.C., Liaw, Y.-C.: Fast-searching algorithm for vector quantization using projection and triangular inequality. IEEE Transactions on Image Processing 13(12), 1554–1558 (2004)CrossRefMathSciNetGoogle Scholar
  18. 18.
    Lampinen, J., Oja, E.: Fast self-organization by the probing algorithm. In: International Joint Conference on Neural Networks, IJCNN, vol. 2, pp. 503–507 (June 1989)Google Scholar
  19. 19.
    Lin, Y.K., Fu, K.S.: Automatic classification of cervical cells using a binary tree classifier. Pattern Recognition 16(1), 69–80 (1983)CrossRefGoogle Scholar
  20. 20.
    Pakkanen, J., Iivarinen, J., Oja, E.: The Evolving Tree — a novel self-organizing network for data analysis. Neural Processing Letters 20(3), 199–211 (2004)CrossRefGoogle Scholar
  21. 21.
    Rahmel, J.: SplitNet: learning of tree structured Kohonen chains. In: IEEE International Conference on Neural Networks, vol. 2, pp. 1221–1226 (June 1996)Google Scholar
  22. 22.
    Rahmel, J., Blum, C., Hahn, P.: On the role of hierarchy for neural network interpretation. In: IJCAI 1997: Proceedings of the Fifteenth International Joint Conference on Artifical Intelligence, pp. 1072–1077. Morgan Kaufmann Publishers Inc., San Francisco (1997)Google Scholar
  23. 23.
    Samet, H.: The quadtree and related hierarchical data structures. ACM Comput. Surv. 16, 187–260 (1984)CrossRefMathSciNetGoogle Scholar
  24. 24.
    Sproull, R.F.: Refinements to nearest-neighbor searching in k-dimensional trees. Algorithmica 6(4), 579–589 (1991)CrossRefzbMATHMathSciNetGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2014

Authors and Affiliations

  • César A. Astudillo
    • 1
  • B. John Oommen
    • 2
  1. 1.Department of Computer ScienceUniversidad de TalcaCuricóChile
  2. 2.School of Computer ScienceCarleton UniversityOttawaCanada

Personalised recommendations