Estimating the Reach of a Manifold

Eddie Aamari (1, 2, 3) , Jisu Kim (4) , Frédéric Chazal (3) , Bertrand Michel (5, 6) , Alessandro Rinaldo (4) , Larry Wasserman (4)

1 LM-Orsay - Laboratoire de Mathématiques d'Orsay
2 SELECT - Model selection in statistical learning
3 DATASHAPE - Understanding the Shape of Data
4 Statistics Department, Carnegie Mellon University
5 ECN - École Centrale de Nantes
6 LMJL - Laboratoire de Mathématiques Jean Leray

Various problems in manifold estimation make use of a quantity called the {\em reach}, denoted by $\tau_M$, which is a measure of the regularity of the manifold. This paper is the first investigation into the problem of how to estimate the reach. First, we study the geometry of the reach through an approximation perspective. We derive new geometric results on the reach for submanifolds without boundary. An estimator $\hat{\tau}$ of $\tau_M$ is proposed in a framework where tangent spaces are known, and bounds assessing its efficiency are derived. In the case of i.i.d. random point cloud $\mathbb{X}_n$, $\hat{\tau}(\mathbb{X}_n)$ is showed to achieve uniform expected loss bounds over a $\mathcal{C}^3$-like model. Finally, we obtain upper and lower bounds on the minimax rate for estimating the reach.

Keywords: Reach; Geometric Inference; Minimax Estimation; Curvature

Domains: Statistics Theory [stat.TH]; Differential Geometry [math.DG]

Eddie Aamari, Jisu Kim, Frédéric Chazal, Bertrand Michel, Alessandro Rinaldo, et al.. Estimating the Reach of a Manifold. 2017. ⟨hal-01521955v1⟩