ML Anthology

A Unifying Approach to Hard and Probabilistic Clustering

Ron Zass, Amnon Shashua
ICCV 2005 pp. 294-301
doi:10.1109/ICCV.2005.27 /iccv/2005/zass2005iccv-unifying/

Abstract

We derive the clustering problem from first principles showing that the goal of achieving a probabilistic, or ”hard”, multi class clustering result is equivalent to the algebraic problem of a completely positive factorization under a doubly stochastic constraint. We show that spectral clustering, normalized cuts, kernel K-means and the various normalizations of the associated affinity matrix are particular instances and approximations of this general principle. We propose an efficient algorithm for achieving a completely positive factorization and extend the basic clustering scheme to situations where partial label information is available. 1.

PDF ICCV Semantic Scholar

Cite

Text

Zass and Shashua. "A Unifying Approach to Hard and Probabilistic Clustering." IEEE/CVF International Conference on Computer Vision, 2005. doi:10.1109/ICCV.2005.27

Markdown

[Zass and Shashua. "A Unifying Approach to Hard and Probabilistic Clustering." IEEE/CVF International Conference on Computer Vision, 2005.](https://mlanthology.org/iccv/2005/zass2005iccv-unifying/) doi:10.1109/ICCV.2005.27

BibTeX

@inproceedings{zass2005iccv-unifying,
  title     = {{A Unifying Approach to Hard and Probabilistic Clustering}},
  author    = {Zass, Ron and Shashua, Amnon},
  booktitle = {IEEE/CVF International Conference on Computer Vision},
  year      = {2005},
  pages     = {294-301},
  doi       = {10.1109/ICCV.2005.27},
  url       = {https://mlanthology.org/iccv/2005/zass2005iccv-unifying/}
}