Advertisement

Enhanced Iterative Projection for Subclass Analysis under EM Framework

  • Yuting Tao
  • Jian Yang
Conference paper
Part of the Communications in Computer and Information Science book series (CCIS, volume 321)

Abstract

Linear discriminant analysis (LDA) is a very popular supervised classification approach. But it cannot perform well in some cases such as large sample size, etc. In terms of its shortcoming, some scholars in this area put up the idea of subclass, which can break out of LDA’s limitation and achieve better classification results. Subclass discriminant analysis (SDA) worked out the division of subclasses, before solving the generalized eigenvalue problem. By contrast, our proposed approach performs subclass division based on K-means cluster, class by class, in the iterative steps under EM framework. The experimental results on two character databases show that our proposed approach can achieve better results than SDA, meanwhile not quite time-consuming.

Keywords

Linear discriminant analysis larger sample size subclass discriminant analysis EM framework iteration generalized eigenvalue problem K-means cluster 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Zhu, M., Martinez, A.M.: Student members, IEEE: Subclass Discriminant Analysis. J. IEEE Trans. Pattern Analysis and Machine Learning 28(8), 1274–1286 (2006)CrossRefGoogle Scholar
  2. 2.
    Fisher, R.A.: The use of multiple measurements in taxonomic problem. J. Annals Eugenics 7(2), 179–188 (1936)Google Scholar
  3. 3.
    Welling, M.: Fisher Linear Discriminant Analysis, http://www.ics.uci.edu/~welling/classnotes/papers_class/Fisher-LDA.pdf
  4. 4.
    Sun, L., Ceran, B., Ye, J.: A Scalable Two-Stage Approach for a Class of Dimensionality Reduction Techniques. In: Proceedings of the 16th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2010, pp. 313–322 (2010)Google Scholar
  5. 5.
    Bishop, C.M.: Pattern recognition and Machine Learning. Springer, New York (2006)zbMATHGoogle Scholar
  6. 6.
    Boyd, S., Vandenberghe, L.: Convex optimization. Cambridge University Press (2004)Google Scholar
  7. 7.
    Sun, L., Ji, S., Ye, J.: Hypergraph Spectral Learning for Multi-label Classification. In: Proceedings of the 14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD 2008, pp. 668–676 (2008)Google Scholar
  8. 8.
    Wang, X., Tang, X.: Dual-Space Linear Discriminant Analysis for Face Recognition. In: Computer Vision and Pattern Recognition, CVPR 2004, vol. 2, pp. 564–569 (2004)Google Scholar
  9. 9.
    Tang, Y.T., et al.: Offline recognition of Chinese handwriting by multi-feature and multi-level classification. IEEE Transactions on Pattern Analysis and Machine Intelligence 20(5), 556–561 (1998)CrossRefGoogle Scholar
  10. 10.
    Tseng, Y.H., Kuo, C.C., Lee, H.J.: Speeding up Chinese character recognition in an automatic document reading system. Pattern Recognition 31(11), 1601–1612 (1998)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Yuting Tao
    • 1
  • Jian Yang
    • 1
  1. 1.Nanjing University of Science and TechnologyNanjingChina

Personalised recommendations