Geodesic least squares regression on the gaussian manifold with an application in astrophysics

Geert Verdoolaege

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

We present a new regression method called geodesic least squares (GLS), which is particularly robust against data and model uncertainty. It is based on minimization of the Rao geodesic distance on a probabilistic manifold. We apply GLS to Tully-Fisher scaling of the total baryonic mass vs. the rotation velocity in disk galaxies and we show the excellent robustness properties of GLS for estimating the coefficients and the tightness of the scaling.

Original languageEnglish
Title of host publicationGeometric Science of Information - 3rd International Conference, GSI 2017, Proceedings
EditorsFrank Nielsen, Frederic Barbaresco, Frank Nielsen
PublisherSpringer
Pages621-628
Number of pages8
ISBN (Print)9783319684444
DOIs
Publication statusPublished - 2017
Event3rd International Conference on Geometric Science of Information, GSI 2017 - Paris, France
Duration: 7 Nov 20179 Nov 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10589 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference3rd International Conference on Geometric Science of Information, GSI 2017
Country/TerritoryFrance
CityParis
Period7/11/179/11/17

Keywords

  • Geodesic least squares
  • Rao geodesic distance
  • Robust regression
  • Tully-Fisher scaling

Fingerprint

Dive into the research topics of 'Geodesic least squares regression on the gaussian manifold with an application in astrophysics'. Together they form a unique fingerprint.

Cite this