Header logo is

Large Margin Non-Linear Embedding

2005

Conference Paper

ei


It is common in classification methods to first place data in a vector space and then learn decision boundaries. We propose reversing that process: for fixed decision boundaries, we ``learn‘‘ the location of the data. This way we (i) do not need a metric (or even stronger structure) -- pairwise dissimilarities suffice; and additionally (ii) produce low-dimensional embeddings that can be analyzed visually. We achieve this by combining an entropy-based embedding method with an entropy-based version of semi-supervised logistic regression. We present results for clustering and semi-supervised classification.

Author(s): Zien, A. and Candela, JQ.
Book Title: ICML 2005
Journal: Proceedings of the 22nd International Conference on Machine Learning (ICML 2005)
Pages: 1065-1072
Year: 2005
Month: August
Day: 0
Editors: De Raedt, L. , S. Wrobel
Publisher: ACM Press

Department(s): Empirical Inference
Bibtex Type: Conference Paper (inproceedings)

DOI: 10.1145/1102351.1102485
Event Name: 22nd International Conference on Machine Learning
Event Place: Bonn, Germany

Address: New York, NY, USA
Digital: 0
Language: en
Organization: Max-Planck-Gesellschaft
School: Biologische Kybernetik

Links: PDF
PostScript
Web

BibTex

@inproceedings{3375,
  title = {Large Margin Non-Linear Embedding},
  author = {Zien, A. and Candela, JQ.},
  journal = {Proceedings of the 22nd International Conference on Machine Learning  (ICML 2005)},
  booktitle = {ICML 2005},
  pages = {1065-1072},
  editors = {De Raedt, L. , S. Wrobel},
  publisher = {ACM Press},
  organization = {Max-Planck-Gesellschaft},
  school = {Biologische Kybernetik},
  address = {New York, NY, USA},
  month = aug,
  year = {2005},
  doi = {10.1145/1102351.1102485},
  month_numeric = {8}
}