Total Variation Regularized Fr\'echet Regression for Metric-Space Valued Data

TL;DR

This paper introduces a total variation regularization method for nonparametric Fréchet regression in metric spaces, enabling effective modeling of complex non-Euclidean data indexed by scalar predictors.

Contribution

It develops a novel total variation regularization technique for Fréchet regression, providing theoretical guarantees and practical algorithms for metric-space valued data.

Findings

Estimator is piece-wise constant and minimax optimal in Hadamard spaces.

Method performs well on simulated data and real-world examples.

Applicable to diverse metric spaces like SPD matrices, distributions, and phylogenetic trees.

Abstract

Non-Euclidean data that are indexed with a scalar predictor such as time are increasingly encountered in data applications, while statistical methodology and theory for such random objects are not well developed yet. To address the need for new methodology in this area, we develop a total variation regularization technique for nonparametric Fr\'echet regression, which refers to a regression setting where a response residing in a metric space is paired with a scalar predictor and the target is a conditional Fr\'echet mean. Specifically, we seek to approximate an unknown metric-space valued function by an estimator that minimizes the Fr\'echet version of least squares and at the same time has small total variation, appropriately defined for metric-space valued objects. We show that the resulting estimator is representable by a piece-wise constant function and establish the minimax convergence rate of the proposed estimator for metric data objects that reside in Hadamard spaces. We illustrate the numerical performance of the proposed method for both simulated and real data, including metric spaces of symmetric positive-definite matrices with the affine-invariant distance, of probability distributions on the real line with the Wasserstein distance, and of phylogenetic trees with the Billera--Holmes--Vogtmann metric.