Advertisement

Constructing Class Hierarchy via Adaptive Kernel Learning

  • Yanting Lu
  • Jianfeng Lu
  • Liantao Wang
  • Jingyu Yang
Part of the Communications in Computer and Information Science book series (CCIS, volume 321)

Abstract

Hierarchical structure, which is universal in nature and human society, can bring much benefit to classification tasks. Hierarchical classification typically first decomposes the multi-class classification problem into many smaller ones according to the hierarchical relationship lying in the classes, and then learns and organizes corresponding classifiers hierarchically. In such a framework, learning the class hierarchy is a critical and challenging step. In this paper, a k-ary class hierarchy construction approach is designed for multi-class and multi-feature scenario. An generalized adaptive kernel learning method is proposed to optimize the kernel fusion and k-way class partition together, and an iterative optimization algorithm is designed for it. Experimental results on synthetic and real datasets show the superiority of k-ary class hierarchy to binary class hierarchy.

Keywords

Class Hierarchy Hierarchical Classification Data Fusion Kernel Learning Centered Kernel Alignment 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. 1.
    Chen, Y., Crawford, M.M., Ghosh, J.: Integrating Support Vector Machines in a Hierarchical Output Space Decomposition Framework. In: IEEE International Geoscience and Remote Sensing Symposium, pp. 949–952. IEEE Press, New York (2004)Google Scholar
  2. 2.
    Casasent, D., Wang, Y.: A Hierarchical Classifier Using New Support Vector Machines for Automatic Target Recognition. Neur. Networks 18, 541–548 (2005)CrossRefGoogle Scholar
  3. 3.
    Marszałek, M., Schmid, C.: Constructing Category Hierarchies for Visual Recognition. In: Forsyth, D., Torr, P., Zisserman, A. (eds.) ECCV 2008, Part IV. LNCS, vol. 5305, pp. 479–491. Springer, Heidelberg (2008)CrossRefGoogle Scholar
  4. 4.
    Cevikalp, H.: New Clustering Algorithms for The Support Vector Machine Based Hierarchical Classification. Pattern Recognition Letters 31, 1285–1291 (2010)CrossRefGoogle Scholar
  5. 5.
    Gao, T., Koller, D.: Discriminative Learning of Relaxed Hierarchy for Large-Scale Visual Recognition. In: IEEE International Conference on Computer Vision, pp. 2072–2079. IEEE Press, New York (2011)Google Scholar
  6. 6.
    Xiao, L., Zhou, D., Wu, M.: Hierarchical Classification via Orthogonal Transfer. In: the 28th International Conference on Machine Learning, pp. 801–808. ACM, New York (2011)Google Scholar
  7. 7.
    Lu, Y., Lu, J., Yang, J.: Adaptive Kernel Learning Based on Centered Alignment for Hierarchical Classification. To appear in 21st International Conference on Pattern Recognition (2012)Google Scholar
  8. 8.
    Marszałek, M., Schmid, C.: Semantic Hierarchies for Visual Object Recognition. In: 25th IEEE Conference on Computer Vision and Pattern Recognition, pp. 2319–2325. IEEE Computer Society, Piscataway (2007)Google Scholar
  9. 9.
    Zweig, A., Weinshall, D.: Exploiting object hierarchy: Combining models from different category levels. In: 11th IEEE International Conference on Computer Vision, pp. 1847–1854. IEEE Computer Society, Piscataway (2007)Google Scholar
  10. 10.
    Yuan, X., Lai, W., Mei, T., Hua, X., Wu, X., Li, S.: Automatic video genre categorization using hierarchical SVM. In: 13th IEEE International Conference on Image Processing, pp. 2905–2908. IEEE Computer Society, Piscataway (2006)Google Scholar
  11. 11.
    Binder, A., Müller, K., Kawanabe, M.: On Taxonomies for Multi-class Image Categorization. International Journal of Computer Vision 99(3), 281–301 (2012)CrossRefGoogle Scholar
  12. 12.
    Griffin, G., Perona, P.: Learning and Using Taxonomies for Fast Visual Categorization. In: 26th IEEE Conference on Computer Vision and Pattern Recognition, pp. 533–540. IEEE Computer Society, Piscataway (2008)Google Scholar
  13. 13.
    Bengio, S., Weston, J., Grangier, D.: Label Embedding Trees for Large Multi-Class Tasks. In: 24th Annual Conference on Neural Information Processing Systems, pp. 163–171. Curran Associates Inc., New York (2010)Google Scholar
  14. 14.
    Deng, J., Satheesh, S., Berg, A.C., Li, F.: Fast and Balanced: Efficient Label Tree Learning for Large Scale Object Recognition. In: 25th Annual Conference on Neural Information Processing Systems, pp. 567–575. Curran Associates Inc., New York (2011)Google Scholar
  15. 15.
    Cortes, C., Mohri, M., Rostamizadeh, A.: Two-stage learning kernel algorithms. In: Proceedings of the 27th International Conference on Machine Learning, pp. 239–246. Omnipress, Haifa (2010)Google Scholar
  16. 16.
    Fan, K.: On a Theorem of Weyl Concerning Eigen-Values of Linear Transformations. In: Proceedings of the National Academy of Sciences of the United States of America, pp. 31–35. National Academy of Sciences, Washington (1950)Google Scholar
  17. 17.
    Zha, H., He, X., Ding, C., Simon, H., Gu, M.: Spectral Relaxation for K-Means Clustering. In: 15th Annual Conference on Neural Information Processing Systems, pp. 1057–1064. MIT Press, Cambridge (2002)Google Scholar
  18. 18.
    Chang, C., Lin, C.: LIBSVM: a library for support vector machines. ACM Trans. Intell. Syst. Technol. 27, 27:1–27:27 (2011)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Yanting Lu
    • 1
  • Jianfeng Lu
    • 1
  • Liantao Wang
    • 1
  • Jingyu Yang
    • 1
  1. 1.School of Computer Science & TechnologyNanjing University of Science & TechnologyNanjingChina

Personalised recommendations