Means in complete manifolds: uniqueness and approximation

DOI10.1051/PS/2013033MaRDI QIDQ5174350zbMATH OpenOpenAlexFDO

Authors Marc Arnaudon, Laurent Miclo

Publication date 17 February 2015

Published in ESAIM: Probability and Statistics (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1207.3232

zbMATH Keywords

diffusion processesGibbs measureshomogenizationinstantaneous invariant measuresintrinsic \(p\)-meansprobability measures on compact Riemannian manifoldssimulated annealingspectral gap at small temperaturestochastic algorithms

Mathematics Subject Classification ID

Methods of global Riemannian geometry, including PDE methods; curvature restrictions (53C21) Integration on manifolds; measures on manifolds (58C35) Geometric probability and stochastic geometry (60D05) Diffusion processes (60J60)

Abstract: Let $M$ be a complete Riemannian manifold, $NinNN$ and $pge1$. We prove that almost everywhere on $x=(x_1,\mathrm{...},x_N)in{M}^N$ for Lebesgue measure in $M^N$, the measure $dimu(x)=f1Nsu{m}_{k=1}^N{d}_{x_k}$ has a unique $p$-mean $e_p(x)$. As a consequence, if $X=(X_1,\mathrm{...},X_N)$ is a $M^N$-valued random variable with absolutely continuous law, then almost surely $mu(X(om))$ has a unique $p$-mean. In particular if $(X_n{)}_{nge1}$ is an independent sample of an absolutely continuous law in $M$, then the process $e_{p,n}(om)=e_p(X_1(om),\mathrm{...},X_n(om))$ is well-defined. Assume $M$ is compact and consider a probability measure $u$ in $M$. Using partial simulated annealing, we define a continuous semimartingale which converges to the set of minimizers of the integral of distance at power $p$ with respect to $u$. When the set is a singleton, it converges to the $p$-mean.

Recommendations

• Uniform approximation on manifolds
• Mean value theorems on manifolds
• Approximation in mean on homogeneous compact spaces
• Uniform algebras and approximation on manifolds
• scientific article; zbMATH DE number 898096
• Uniqueness of the best approximation in mean of vector-valued functions
• Iterations of the mean-type mappings and uniqueness of invariant means
• Unimodular measures on the space of all Riemannian manifolds
• Uniqueness theorems for measures in \(L_ r\) and \(C_ o(\Omega)\)
• On the convergence of Riesz means on compact manifolds

Cited in

(15)

• Distances and means of direct similarities
• Principal component analysis and clustering on manifolds
• Geometrical smeariness -- a new phenomenon of Fréchet means
• A smeary central limit theorem for manifolds with application to high-dimensional spheres
• An algorithm for computing Fréchet means on the sphere
• Riemannian \(L^{p}\) center of mass: existence, uniqueness, and convexity
• Riemannian Means as Solutions of Variational Problems
• Stochastic algorithms for computing means of probability measures
• How to calculate the barycenter of a weighted graph
• On the measure of the cut locus of a Fréchet mean
• A hyperbolic approach for learning communities on graphs
• Geometry of sample spaces
• Finite sample smeariness of Fréchet means with application to climate
• Data analysis on nonstandard spaces
• High-dimensional Hilbert-Schmidt linear regression with Hilbert manifold variables

This page was built for publication: Means in complete manifolds: uniqueness and approximation

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5174350)