Exploring Multiple Communities with Kernel-Based Link Analysis

  • Takahiko Ito
  • Masashi Shimbo
  • Daichi Mochihashi
  • Yuji Matsumoto
Part of the Lecture Notes in Computer Science book series (LNCS, volume 4213)


We discuss issues raised by applying von Neumann kernels to graphs with multiple communities. Depending on the parameter setting, Kandola et al.’s von Neumann kernels can identify not only nodes related to a given node but also the most important nodes in a graph. However, when von Neumann kernels are biased towards importance, top-ranked nodes are the important nodes in the dominant community of the graph irrespective of the communities where the target node belongs. To solve this “topic-drift” problem, we apply von Neumann kernels to the weighted graphs (community graph), which are derived from a generative model of links.


Latent Dirichlet Allocation Citation Network Query Term Community Graph Discourse Processing 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


  1. 1.
    Kleinberg, J.M.: Authoritative sources in a hyperlinked environment. J. ACM, 604–632 (1999)Google Scholar
  2. 2.
    Small, H.: Co-citation in the scientific literature: a new measure of the relationship between two documents. J. American Society for Information Science 24, 265–269 (1973)CrossRefGoogle Scholar
  3. 3.
    Ito, T., Shimbo, M., Kudo, T., Matsumoto, Y.: Application of kernels to link analysis. In: Proc. 11th ACM SIGKDD, pp. 586–592 (2005)Google Scholar
  4. 4.
    Kandola, J., Shawe-Taylor, J., Cristianini, N.: Learning semantic similarity. NIPS 15 (2002)Google Scholar
  5. 5.
    Bharat, K., Henzinger, M.R.: Improved algorithms for topic distillation in a hyperlinked enviornment. In: Proc. 21st ACM SIGIR Conference (1998)Google Scholar
  6. 6.
    Cohn, D., Chang, H.: Learning to probabilistically identify authoritative documents. In: Proc. 18th International Conference of Machine Learning (2001)Google Scholar
  7. 7.
    Hofmann, T.: Learning the similarity of documents: An information-geometric approach to document retrieval and categorization. NIPS 12, 914–920 (2000)Google Scholar
  8. 8.
    Shimbo, M., Ito, T.: Kernels as link analysis measures. In: Cook, D., Holder, L. (eds.) Mining Graph Data. John Wiley & Sons, Chichester (2006)Google Scholar
  9. 9.
    Hofmann, T.: Probabilistic latent semantic indexing. In: Proc. 22th ACM SIGIR Conference, pp. 50–57 (1999)Google Scholar
  10. 10.
    Haussler, D.: Convolution kernels on discrete structures. Technical Report UCSC-CRL-99-10, University of California at Santa Cruz (1999)Google Scholar
  11. 11.
    Jaakkola, T., Haussler, D.: Exploiting generative models in discriminative classifiers. In: NIPS, vol. 11 (1998)Google Scholar
  12. 12.
    Blei, D.M., Ng, A.Y., Jordan, M.I.: Latent Dirichlet Allocation. NIPS 14 (2001)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2006

Authors and Affiliations

  • Takahiko Ito
    • 1
  • Masashi Shimbo
    • 1
  • Daichi Mochihashi
    • 2
  • Yuji Matsumoto
    • 1
  1. 1.Graduate School of Information Science Nara Institute of Science and Technology 
  2. 2.ATR Spoken Language Communication Research Laboratories 

Personalised recommendations