Cluster Trees on Manifolds

Abstract

We investigate the problem of estimating the cluster tree for a density $f$ supported on or near a smooth $d$-dimensional manifold $M$ isometrically embedded in $\mathbb{R}^D$. We study a $k$-nearest neighbor based algorithm recently proposed by Chaudhuri and Dasgupta. Under mild assumptions on $f$ and $M$, we obtain rates of convergence that depend on $d$ only but not on the ambient dimension $D$. We also provide a sample complexity lower bound for a natural class of clustering algorithms that use $D$-dimensional neighborhoods.

Cite

Text

Balakrishnan et al. "Cluster Trees on Manifolds." Neural Information Processing Systems, 2013.

Markdown

[Balakrishnan et al. "Cluster Trees on Manifolds." Neural Information Processing Systems, 2013.](https://mlanthology.org/neurips/2013/balakrishnan2013neurips-cluster/)

BibTeX

@inproceedings{balakrishnan2013neurips-cluster,
  title     = {{Cluster Trees on Manifolds}},
  author    = {Balakrishnan, Sivaraman and Narayanan, Srivatsan and Rinaldo, Alessandro and Singh, Aarti and Wasserman, Larry},
  booktitle = {Neural Information Processing Systems},
  year      = {2013},
  pages     = {2679-2687},
  url       = {https://mlanthology.org/neurips/2013/balakrishnan2013neurips-cluster/}
}