English

Alpha shapes in kernel density estimation

Algebraic Topology 2024-05-02 v3

Abstract

For every Gaussian kernel density estimator f(x)=iaiexp(xxi2/2h2)f(x)=\sum_i a_i \exp(-\lVert x-x_i\rVert^2/2h^2) associated to a point cloud D={x1,...,xN}Rd\mathcal{D}=\{x_1,...,x_N\}\subset \mathbb{R}^d, we define a nested family of closed subspaces S(a)Rd\mathcal{S}(a)\subset\mathbb{R}^d, which we interpret as a continuous version of an alpha shape. Using arguments based on Fenchel duality, we prove that S(a)\mathcal{S}(a) is homotopy equivalent to the superlevel set L(a)=f1[ea,)\mathcal{L}(a)=f^{-1}[e^{-a},\infty), and that L(a)\mathcal{L}(a) can be realized as the union of a certain power-shifted covering by balls with centers in S(a)\mathcal{S}(a). By extracting finite alpha complexes with vertices in S(a)\mathcal{S}(a), we obtain refined geometric models of noisy point clouds, as well as density-filtered persistent homology calculations. In order to compute alpha complexes in higher dimension, we used a recent algorithm due to the present authors based on the duality principle.

Keywords

Cite

@article{arxiv.2303.12213,
  title  = {Alpha shapes in kernel density estimation},
  author = {Erik Carlsson and John Carlsson},
  journal= {arXiv preprint arXiv:2303.12213},
  year   = {2024}
}

Comments

19 pages

R2 v1 2026-06-28T09:27:26.361Z