Modelling anisotropic covariance using stochastic development and sub-Riemannian frame bundle geometry

Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt

Standard

Modelling anisotropic covariance using stochastic development and sub-Riemannian frame bundle geometry. / Sommer, Stefan Horst; Svane, Anne Marie.

I: Journal of Geometric Mechanics, Bind 9, Nr. 3, 2017, s. 391-410.

Publikation: Bidrag til tidsskrift › Tidsskriftartikel › Forskning › fagfællebedømt

Harvard

Sommer, SH & Svane, AM 2017, 'Modelling anisotropic covariance using stochastic development and sub-Riemannian frame bundle geometry', Journal of Geometric Mechanics, bind 9, nr. 3, s. 391-410. https://doi.org/10.3934/jgm.2017015

APA

Sommer, S. H., & Svane, A. M. (2017). Modelling anisotropic covariance using stochastic development and sub-Riemannian frame bundle geometry. Journal of Geometric Mechanics, 9(3), 391-410. https://doi.org/10.3934/jgm.2017015

Vancouver

Sommer SH, Svane AM. Modelling anisotropic covariance using stochastic development and sub-Riemannian frame bundle geometry. Journal of Geometric Mechanics. 2017;9(3):391-410. https://doi.org/10.3934/jgm.2017015

Author

Sommer, Stefan Horst ; Svane, Anne Marie. / Modelling anisotropic covariance using stochastic development and sub-Riemannian frame bundle geometry. I: Journal of Geometric Mechanics. 2017 ; Bind 9, Nr. 3. s. 391-410.

Bibtex

@article{323bbff18e3e4fbab06e6a19b489c518,
title = "Modelling anisotropic covariance using stochastic development and sub-Riemannian frame bundle geometry",
abstract = "We discuss the geometric foundation behind the use of stochastic processes in the frame bundle of a smooth manifold to build stochastic models with applications in statistical analysis of non-linear data. The transition densities for the projection to the manifold of Brownian motions developed in the frame bundle lead to a family of probability distributions on the manifold. We explain how data mean and covariance can be interpreted as points in the frame bundle or, more precisely, in the bundle of symmetric positive definite 2-tensors analogously to the parameters describing Euclidean normal distributions. We discuss a factorization of the frame bundle projection map through this bundle, the natural sub-Riemannian structure of the frame bundle, the effect of holonomy, and the existence of subbundles where the Hormander condition is satisfied such that the Brownian motions have smooth transition densities. We identify the most probable paths for the underlying Euclidean Brownian motion and discuss small time asymptotics of the transition densities on the manifold. The geometric setup yields an intrinsic approach to the estimation of mean and covariance in non-linear spaces.",
keywords = "math.DG, math.ST, stat.TH",
author = "Sommer, {Stefan Horst} and Svane, {Anne Marie}",
year = "2017",
doi = "10.3934/jgm.2017015",
language = "English",
volume = "9",
pages = "391--410",
journal = "Journal of Geometric Mechanics",
issn = "1941-4889",
publisher = "American Institute of Mathematical Sciences",
number = "3",

}

RIS

TY - JOUR

T1 - Modelling anisotropic covariance using stochastic development and sub-Riemannian frame bundle geometry

AU - Sommer, Stefan Horst

AU - Svane, Anne Marie

PY - 2017

Y1 - 2017

N2 - We discuss the geometric foundation behind the use of stochastic processes in the frame bundle of a smooth manifold to build stochastic models with applications in statistical analysis of non-linear data. The transition densities for the projection to the manifold of Brownian motions developed in the frame bundle lead to a family of probability distributions on the manifold. We explain how data mean and covariance can be interpreted as points in the frame bundle or, more precisely, in the bundle of symmetric positive definite 2-tensors analogously to the parameters describing Euclidean normal distributions. We discuss a factorization of the frame bundle projection map through this bundle, the natural sub-Riemannian structure of the frame bundle, the effect of holonomy, and the existence of subbundles where the Hormander condition is satisfied such that the Brownian motions have smooth transition densities. We identify the most probable paths for the underlying Euclidean Brownian motion and discuss small time asymptotics of the transition densities on the manifold. The geometric setup yields an intrinsic approach to the estimation of mean and covariance in non-linear spaces.

AB - We discuss the geometric foundation behind the use of stochastic processes in the frame bundle of a smooth manifold to build stochastic models with applications in statistical analysis of non-linear data. The transition densities for the projection to the manifold of Brownian motions developed in the frame bundle lead to a family of probability distributions on the manifold. We explain how data mean and covariance can be interpreted as points in the frame bundle or, more precisely, in the bundle of symmetric positive definite 2-tensors analogously to the parameters describing Euclidean normal distributions. We discuss a factorization of the frame bundle projection map through this bundle, the natural sub-Riemannian structure of the frame bundle, the effect of holonomy, and the existence of subbundles where the Hormander condition is satisfied such that the Brownian motions have smooth transition densities. We identify the most probable paths for the underlying Euclidean Brownian motion and discuss small time asymptotics of the transition densities on the manifold. The geometric setup yields an intrinsic approach to the estimation of mean and covariance in non-linear spaces.

KW - math.DG

KW - math.ST

KW - stat.TH

U2 - 10.3934/jgm.2017015

DO - 10.3934/jgm.2017015

M3 - Journal article

VL - 9

SP - 391

EP - 410

JO - Journal of Geometric Mechanics

JF - Journal of Geometric Mechanics

SN - 1941-4889

IS - 3

ER -

ID: 179586584