Distance-based Depths for Directional Data

TL;DR

This paper introduces a new class of distance-based depth functions for directional data on spheres, offering computational efficiency in high dimensions and demonstrating advantages in location estimation and classification tasks.

Contribution

It proposes a flexible, computationally efficient class of depth functions for directional data, improving over existing methods especially in high-dimensional settings.

Findings

Depth functions are computationally feasible in high dimensions.

Distance-based depths outperform competitors in simulations.

Proposed depths are robust and have desirable asymptotic properties.

Abstract

Directional data are constrained to lie on the unit sphere of~$\mathbb{R}^q$ for some~$q\geq 2$. To address the lack of a natural ordering for such data, depth functions have been defined on spheres. However, the depths available either lack flexibility or are so computationally expensive that they can only be used for very small dimensions~$q$. In this work, we improve on this by introducing a class of distance-based depths for directional data. Irrespective of the distance adopted, these depths can easily be computed in high dimensions too. We derive the main structural properties of the proposed depths and study how they depend on the distance used. We discuss the asymptotic and robustness properties of the corresponding deepest points. We show the practical relevance of the proposed depths in two applications, related to (i) spherical location estimation and (ii) supervised classification. For both problems, we show through simulation studies that distance-based depths have strong advantages over their competitors.