English

Helly-Type Theorems in Property Testing

Computational Geometry 2013-12-17 v2

Abstract

Helly's theorem is a fundamental result in discrete geometry, describing the ways in which convex sets intersect with each other. If SS is a set of nn points in RdR^d, we say that SS is (k,G)(k,G)-clusterable if it can be partitioned into kk clusters (subsets) such that each cluster can be contained in a translated copy of a geometric object GG. In this paper, as an application of Helly's theorem, by taking a constant size sample from SS, we present a testing algorithm for (k,G)(k,G)-clustering, i.e., to distinguish between two cases: when SS is (k,G)(k,G)-clusterable, and when it is ϵ\epsilon-far from being (k,G)(k,G)-clusterable. A set SS is ϵ\epsilon-far (0<ϵ1)(0<\epsilon\leq1) from being (k,G)(k,G)-clusterable if at least ϵn\epsilon n points need to be removed from SS to make it (k,G)(k,G)-clusterable. We solve this problem for k=1k=1 and when GG is a symmetric convex object. For k>1k>1, we solve a weaker version of this problem. Finally, as an application of our testing result, in clustering with outliers, we show that one can find the approximate clusters by querying a constant size sample, with high probability.

Keywords

Cite

@article{arxiv.1307.8268,
  title  = {Helly-Type Theorems in Property Testing},
  author = {Sourav Chakraborty and Rameshwar Pratap and Sasanka Roy and Shubhangi Saraf},
  journal= {arXiv preprint arXiv:1307.8268},
  year   = {2013}
}
R2 v1 2026-06-22T01:01:17.914Z