Advanced search
1 file | 891.74 KB

Distance dependent extensions of the Chinese restaurant process

Author
Organization
Abstract
In this paper we consider the clustering of text documents using the Chinese Restau- rant Process (CRP) and extensions that take time-correlations into account. To this pur- pose, we implement and test the Distance Dependent Chinese Restaurant Process (DD- CRP) for mixture models on both generated and real-world data. We also propose and im- plement a novel clustering algorithm, the Av- eraged Distance Dependent Chinese Restau- rant Process (ADDCRP), to model time- correlations, that is faster per iteration and attains similar performance as the fully dis- tance dependent CRP.
Keywords
Nonparametric methods, Dirichlet Process, Machine learning

Downloads

  • paper temp.pdf
    • full text
    • |
    • open access
    • |
    • PDF
    • |
    • 891.74 KB

Citation

Please use this url to cite or link to this publication:

Chicago
Wybo, Willem, Camille Colle, Pieter-Jan Kindermans, and Benjamin Schrauwen. 2012. “Distance Dependent Extensions of the Chinese Restaurant Process.” In Proceedings of the 21st Belgian-Dutch Conference on Machine Learning, ed. Bernard De Baets, Bernard Manderick, Michaël Rademaker, and Willem Waegeman. Ghent, Belgium: University Press.
APA
Wybo, W., Colle, C., Kindermans, P.-J., & Schrauwen, B. (2012). Distance dependent extensions of the Chinese restaurant process. In Bernard De Baets, B. Manderick, M. Rademaker, & W. Waegeman (Eds.), Proceedings of the 21st Belgian-Dutch Conference on Machine Learning. Presented at the 21st Annual Belgian-Dutch conference on Machine Learning (BeNeLearn & PMLS 2012), Ghent, Belgium: University Press.
Vancouver
1.
Wybo W, Colle C, Kindermans P-J, Schrauwen B. Distance dependent extensions of the Chinese restaurant process. In: De Baets B, Manderick B, Rademaker M, Waegeman W, editors. Proceedings of the 21st Belgian-Dutch Conference on Machine Learning. Ghent, Belgium: University Press; 2012.
MLA
Wybo, Willem, Camille Colle, Pieter-Jan Kindermans, et al. “Distance Dependent Extensions of the Chinese Restaurant Process.” Proceedings of the 21st Belgian-Dutch Conference on Machine Learning. Ed. Bernard De Baets et al. Ghent, Belgium: University Press, 2012. Print.
@inproceedings{2134615,
  abstract     = {In this paper we consider the clustering of text documents using the Chinese Restau- rant Process (CRP) and extensions that take time-correlations into account. To this pur- pose, we implement and test the Distance Dependent Chinese Restaurant Process (DD- CRP) for mixture models on both generated and real-world data. We also propose and im- plement a novel clustering algorithm, the Av- eraged Distance Dependent Chinese Restau- rant Process (ADDCRP), to model time- correlations, that is faster per iteration and attains similar performance as the fully dis- tance dependent CRP.},
  author       = {Wybo, Willem and Colle, Camille and Kindermans, Pieter-Jan and Schrauwen, Benjamin},
  booktitle    = {Proceedings of the 21st Belgian-Dutch Conference on Machine Learning},
  editor       = {De Baets, Bernard and Manderick, Bernard  and Rademaker, Micha{\"e}l and Waegeman, Willem},
  isbn         = {9789461970442},
  keyword      = {Nonparametric methods,Dirichlet Process,Machine learning},
  language     = {eng},
  location     = {Ghent, Belgium},
  pages        = {6},
  publisher    = {University Press},
  title        = {Distance dependent extensions of the Chinese restaurant process},
  year         = {2012},
}