English

A bi-criteria approximation algorithm for $k$ Means

Data Structures and Algorithms 2015-08-04 v2

Abstract

We consider the classical kk-means clustering problem in the setting bi-criteria approximation, in which an algoithm is allowed to output βk>k\beta k > k clusters, and must produce a clustering with cost at most α\alpha times the to the cost of the optimal set of kk clusters. We argue that this approach is natural in many settings, for which the exact number of clusters is a priori unknown, or unimportant up to a constant factor. We give new bi-criteria approximation algorithms, based on linear programming and local search, respectively, which attain a guarantee α(β)\alpha(\beta) depending on the number βk\beta k of clusters that may be opened. Our gurantee α(β)\alpha(\beta) is always at most 9+ϵ9 + \epsilon and improves rapidly with β\beta (for example: α(2)<2.59\alpha(2)<2.59, and α(3)<1.4\alpha(3) < 1.4). Moreover, our algorithms have only polynomial dependence on the dimension of the input data, and so are applicable in high-dimensional settings.

Keywords

Cite

@article{arxiv.1507.04227,
  title  = {A bi-criteria approximation algorithm for $k$ Means},
  author = {Konstantin Makarychev and Yury Makarychev and Maxim Sviridenko and Justin Ward},
  journal= {arXiv preprint arXiv:1507.04227},
  year   = {2015}
}
R2 v1 2026-06-22T10:12:23.682Z