Nonnegative Lagrangian Relaxation of K-Means and Spectral Clustering

Abstract

We show that K-means and spectral clustering objective functions can be written as a trace of quadratic forms. Instead of relaxation by eigenvectors, we propose a novel relaxation maintaining the nonnegativity of the cluster indicators and thus give the cluster posterior probabilities, therefore resolving cluster assignment difficulty in spectral relaxation. We derive a multiplicative updating algorithm to solve the nonnegative relaxation problem. The method is briefly extended to semi-supervised classification and semi-supervised clustering.

Cite

Text

Ding et al. "Nonnegative Lagrangian Relaxation of K-Means and Spectral Clustering." European Conference on Machine Learning, 2005. doi:10.1007/11564096_51

Markdown

[Ding et al. "Nonnegative Lagrangian Relaxation of K-Means and Spectral Clustering." European Conference on Machine Learning, 2005.](https://mlanthology.org/ecmlpkdd/2005/ding2005ecml-nonnegative/) doi:10.1007/11564096_51

BibTeX

@inproceedings{ding2005ecml-nonnegative,
  title     = {{Nonnegative Lagrangian Relaxation of K-Means and Spectral Clustering}},
  author    = {Ding, Chris H. Q. and He, Xiaofeng and Simon, Horst D.},
  booktitle = {European Conference on Machine Learning},
  year      = {2005},
  pages     = {530-538},
  doi       = {10.1007/11564096_51},
  url       = {https://mlanthology.org/ecmlpkdd/2005/ding2005ecml-nonnegative/}
}