Empirical Inference Conference Paper 2005

Active Learning for Parzen Window Classifier

The problem of active learning is approached in this paper by minimizing directly an estimate of the expected test error. The main difficulty in this ``optimal'' strategy is that output probabilities need to be estimated accurately. We suggest here different methods for estimating those efficiently. In this context, the Parzen window classifier is considered because it is both simple and probabilistic. The analysis of experimental results highlights that regularization is a key ingredient for this strategy.

Author(s): Chapelle, O.
Book Title: AISTATS 2005
Journal: AI STATS
Pages: 49-56
Year: 2005
Month: January
Day: 0
Editors: Cowell, R. , Z. Ghahramani
Bibtex Type: Conference Paper (inproceedings)
Event Name: Tenth International Workshop on Artificial Intelligence and Statistics (AI & Statistics 2005)
Event Place: Barbados
Digital: 0
Electronic Archiving: grant_archive
ISBN: 0-9727358-1-X
Organization: Max-Planck-Gesellschaft
School: Biologische Kybernetik
Links:

BibTex

@inproceedings{2563,
  title = {Active Learning for Parzen Window Classifier},
  journal = {AI STATS},
  booktitle = {AISTATS 2005},
  abstract = {The problem of active learning is approached in this paper by minimizing
  directly an estimate of the expected test error. The main difficulty
  in this ``optimal'' strategy is that output probabilities need to be 
  estimated accurately. We suggest here different methods
  for estimating those efficiently.
  In this context, the Parzen window classifier is considered
  because it is both simple and probabilistic. The analysis of experimental
  results highlights that regularization is a key ingredient for this strategy.},
  pages = {49-56},
  editors = {Cowell, R. , Z. Ghahramani},
  organization = {Max-Planck-Gesellschaft},
  school = {Biologische Kybernetik},
  month = jan,
  year = {2005},
  slug = {2563},
  author = {Chapelle, O.},
  month_numeric = {1}
}