Fréchet regression for random objects with Euclidean predictors

DOI10.1214/17-AOS1624zbMATH Open1417.62091arXiv1608.03012OpenAlexW2963448920MaRDI QIDQ132099FDOQ132099

Hans-Georg Müller, Hans-Georg Müller, Alexander Petersen, Alexander Petersen

Publication date: 1 April 2019

Published in: The Annals of Statistics (Search for Journal in Brave)

Abstract: Increasingly, statisticians are faced with the task of analyzing complex data that are non-Euclidean and specifically do not lie in a vector space. To address the need for statistical methods for such data, we introduce the concept of Fr'echet regression. This is a general approach to regression when responses are complex random objects in a metric space and predictors are in $mathcal{R}^p$, achieved by extending the classical concept of a Fr'echet mean to the notion of a conditional Fr'echet mean. We develop generalized versions of both global least squares regression and local weighted least squares smoothing. The target quantities are appropriately defined population versions of global and local regression for response objects in a metric space. We derive asymptotic rates of convergence for the corresponding fitted regressions using observed data to the population targets under suitable regularity conditions by applying empirical process methods. For the special case of random objects that reside in a Hilbert space, such as regression models with vector predictors and functional data as responses, we obtain a limit distribution. The proposed methods have broad applicability. Illustrative examples include responses that consist of probability distributions and correlation matrices, and we demonstrate both global and local Fr'echet regression for demographic and brain imaging data. Local Fr'echet regression is also illustrated via a simulation with response data which lie on the sphere.

Full work available at URL: https://arxiv.org/abs/1608.03012

zbMATH Keywords

local linear regressionbrain imaginedensities as objectsfunctional connectivityleast squares regressionmetric spacesrandom objectsWasserstein space

Mathematics Subject Classification ID

Nonparametric estimation (62G05) Nonparametric regression and quantile regression (62G08) Asymptotic properties of nonparametric inference (62G20) Applications of statistics to biology and medical sciences; meta analysis (62P10) Image analysis in multivariate analysis (62H35)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Manopt, a Matlab toolbox for optimization on manifolds
• Computing the nearest correlation matrix--a problem from finance
• Flexible smoothing with \(B\)-splines and penalties. With comments and a rejoinder by the authors
• Weak convergence and empirical processes. With applications to statistics
• Regression Models on Riemannian Symmetric Spaces
• Smoothing noisy data with spline functions: Estimating the correct degree of smoothing by the method of generalized cross-validation
• Statistical Analysis of Spherical Data
• Fréchet regression for random objects with Euclidean predictors
• Regression Analysis for a Functional Response
• Geodesic regression and the theory of least squares on Riemannian manifolds
• A Quadratically Convergent Newton Method for Computing the Nearest Correlation Matrix
• Nonparametric Regression between General Riemannian Manifolds
• Spherical regression with errors in variables
• Large sample theory of intrinsic and extrinsic sample means on manifolds. I
• Object oriented data analysis: sets of trees
• Nonparametric statistics on manifolds and their applications to object data analysis
• Amplitude and phase variation of point processes
• Distances and inference for covariance operators
• Overview of object oriented data analysis
• Riemannian $L^{p}$ center of mass: Existence, uniqueness, and convexity
• Central limit theorems for Fréchet means in the space of phylogenetic trees
• Regression for non-Euclidean data using distance matrices
• A preconditioned Newton algorithm for the nearest correlation matrix
• On power and sample size determinations for the Wilcoxon–Mann–Whitney test
• Geometric Means in a Novel Vector Space Structure on Symmetric Positive‐Definite Matrices
• On the measure of the cut locus of a Fréchet mean
• Local polynomial regression for symmetric positive definite matrices
• Wasserstein geometry of Gaussian measures

Cited In (53)

• Exponential concentration for geometric-median-of-means in non-positive curvature spaces
• Dimension Reduction for Fréchet Regression
• Computing regularized splines in the Riemannian manifold of probability measures
• Wasserstein Regression
• Fréchet single index models for object response regression
• Modeling sparse longitudinal data on Riemannian manifolds
• Intrinsic Riemannian functional data analysis for sparse longitudinal observations
• Convergence rates for the generalized Fréchet mean via the quadruple inequality
• Functional Additive Models on Manifolds of Planar Shapes and Forms
• Total variation regularized Fréchet regression for metric-space valued data
• The Stein effect for Fréchet means
• Uniform convergence of local Fréchet regression with applications to locating extrema and time warping for metric space valued trajectories
• Additive Functional Regression for Densities as Responses
• Random distributional response model based on spline method
• Discussion: A Spatial Modeling Approach for Linguistic Object Data: Analyzing Dialect Sound Variations Across Great Britain, by Shahin Tavakoli et al.
• Rejoinder for “A Spatial Modeling Approach for Linguistic Object Data: Analyzing Dialect Sound Variations Across Great Britain”
• Fisher-Pitman Permutation Tests Based on Nonparametric Poisson Mixtures with Application to Single Cell Genomics
• Rejoinder on: ``Recent advances in directional statistics
• Comments on: ``Recent advances in directional statistics
• Fréchet change-point detection
• Wasserstein \(F\)-tests and confidence bands for the Fréchet regression of density response curves
• Additive regression for non-Euclidean responses and predictors
• From multivariate to functional data analysis: fundamentals, recent developments, and emerging areas
• The effect of correlated errors on the performance of local linear estimation of regression function based on random functional design
• Wasserstein gradients for the temporal evolution of probability distributions
• Latent deformation models for multivariate functional data and time-warping separability
• Nonparametric regression in nonstandard spaces
• Fréchet kernel sliced inverse regression
• Variable Selection for Global Fréchet Regression
• Distribution-on-distribution regression with Wasserstein metric: multivariate Gaussian case
• Intrinsic Riemannian functional data analysis
• On Multiscale Quasi-Interpolation of Scattered Scalar- and Manifold-Valued Functions
• Modeling Time-Varying Random Objects and Dynamic Networks
• Single index Fréchet regression
• Title not available (Why is that?)
• Medoid splits for efficient random forests in metric spaces
• Concurrent object regression
• Numerical algorithms for spline interpolation on space of probability density functions
• Graph-valued regression: prediction of unlabelled networks in a non-Euclidean graph space
• Functional Outlier Detection for Density-Valued Data with Application to Robustify Distribution-to-Distribution Regression
• Metrics and barycenters for point pattern data
• Two-sample and change-point inference for non-Euclidean valued time series
• Sliced Inverse Regression in Metric Spaces
• Point process models for COVID-19 cases and deaths
• A Spatial Modeling Approach for Linguistic Object Data: Analyzing Dialect Sound Variations Across Great Britain
• Rank dynamics for functional data
• Gaussian Process Subspace Prediction for Model Reduction
• frechet
• Representation and reconstruction of covariance operators in linear inverse problems
• Decomposition and reproducing property of local polynomial equivalent kernels in varying coefficient models
• Metric statistics: exploration and inference for random objects with distance profiles
• Data analysis on nonstandard spaces
• Fréchet regression for random objects with Euclidean predictors

Uses Software

• Manopt

Recommendations

• Fréchet analysis of variance for random objects 👍 👎
• Fréchet means and Procrustes analysis in Wasserstein space 👍 👎
• Total variation regularized Fréchet regression for metric-space valued data 👍 👎
• Nonparametric regression in nonstandard spaces 👍 👎
• The Stein effect for Fréchet means 👍 👎
• Uniform convergence of local Fréchet regression with applications to locating extrema and time warping for metric space valued trajectories 👍 👎
• Nonparametric analysis of non-Euclidean data on shapes and images 👍 👎
• Distribution-on-distribution regression via optimal transport maps 👍 👎
• Wasserstein Regression 👍 👎
• Fréchet single index models for object response regression 👍 👎

This page was built for publication: Fréchet regression for random objects with Euclidean predictors

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q132099)