Adaptive estimation of convex and polytopal support

TL;DR

This paper introduces adaptive estimators for the support of a uniform density assuming it is convex or polytopal, achieving dimension-independent rates and improved risk bounds, especially in higher dimensions.

Contribution

It presents a novel estimator for convex and polytopal supports with dimension-independent rates and near-minimax optimality, improving upon previous methods.

Findings

Estimator achieves dimension-independent convergence rates.

For dimensions d≥3, the estimator outperforms previous methods.

Proposes an adaptive estimator sensitive to boundary structure.

Abstract

We estimate the support of a uniform density, when it is assumed to be a convex polytope or, more generally, a convex body in $\R^d$. In the polytopal case, we construct an estimator achieving a rate which does not depend on the dimension $d$, unlike the other estimators that have been proposed so far. For $d\geq 3$, our estimator has a better risk than the previous ones, and it is nearly minimax, up to a logarithmic factor. We also propose an estimator which is adaptive with respect to the structure of the boundary of the unknown support.