Geodesic PCA versus Log-PCA of Histograms in the Wasserstein Space
From MaRDI portal
Publication:4610140
DOI10.1137/17M1143459zbMATH Open1387.62077arXiv1708.08143OpenAlexW2795320988WikidataQ125801157 ScholiaQ125801157MaRDI QIDQ4610140FDOQ4610140
Authors: Elsa Cazelles, Vivien Seguy, Jérémie Bigot, Marco Cuturi, Nicolas Papadakis 
Publication date: 5 April 2018
Published in: SIAM Journal on Scientific Computing (Search for Journal in Brave)
Abstract: This paper is concerned by the statistical analysis of data sets whose elements are random histograms. For the purpose of learning principal modes of variation from such data, we consider the issue of computing the PCA of histograms with respect to the 2-Wasserstein distance between probability measures. To this end, we propose to compare the methods of log-PCA and geodesic PCA in the Wasserstein space as introduced by Bigot et al. (2015) and Seguy and Cuturi (2015). Geodesic PCA involves solving a non-convex optimization problem. To solve it approximately, we propose a novel forward-backward algorithm. This allows a detailed comparison between log-PCA and geodesic PCA of one-dimensional histograms, which we carry out using various data sets, and stress the benefits and drawbacks of each method. We extend these results for two-dimensional data and compare both methods in that setting.
Full work available at URL: https://arxiv.org/abs/1708.08143
Recommendations
- Geodesic PCA in the Wasserstein space by convex PCA
- Wasserstein Riemannian geometry of Gaussian densities
- Geodesic distances in the maximum likelihood estimator of intrinsic dimensionality
- Deconvolution for the Wasserstein metric and geometric inference
- Deconvolution for the Wasserstein metric and geometric inference
- Wasserstein Riemannian geometry on statistical manifold
- Bures–Wasserstein Minimizing Geodesics between Covariance Matrices of Different Ranks
- Wasserstein geometry of Gaussian measures
- Gromov-Wasserstein distances between Gaussian distributions
- Geodesic Convexity and Covariance Estimation
Factor analysis and principal components; correspondence analysis (62H25) Numerical optimization and variational techniques (65K10) Convex programming (90C25)
Cites Work
- iPiano: inertial proximal algorithm for nonconvex optimization
- Title not available (Why is that?)
- Optimal entropy-transport problems and a new Hellinger-Kantorovich distance between positive measures
- Barycenters in the Wasserstein space
- Polar factorization and monotone rearrangement of vector‐valued functions
- Geodesic regression and the theory of least squares on Riemannian manifolds
- Functional data analysis for density functions by transformation to a Hilbert space
- Measure theory and fine properties of functions
- Title not available (Why is that?)
- Title not available (Why is that?)
- Convergence of descent methods for semi-algebraic and tame problems: proximal algorithms, forward-backward splitting, and regularized Gauss-Seidel methods
- Gradient flows with metric and differentiable structures, and applications to the Wasserstein space
- Iterative Bregman projections for regularized transportation problems
- An interpolating distance between optimal transport and Fisher-Rao metrics
- On the ergodic convergence rates of a first-order primal-dual algorithm
- An inertial forward-backward algorithm for monotone inclusions
- A linear optimal transportation framework for quantifying and visualizing variations in sets of images
- Geodesic PCA in the Wasserstein space by convex PCA
- Unbalanced optimal transport: dynamic and Kantorovich formulations
- Geometric Methods for Spectral Analysis
Cited In (23)
- Wasserstein Regression
- Wasserstein discriminant analysis
- Fréchet change-point detection
- Semi-discrete optimal transport: hardness, regularization and numerical solution
- Wasserstein gradients for the temporal evolution of probability distributions
- Long-time principal geodesic analysis in director-based dynamics of hybrid mechanical systems
- Title not available (Why is that?)
- Projection-based techniques for high-dimensional optimal transport problems
- Wasserstein principal component analysis for circular measures
- Quantitative stability of optimal transport maps under variations of the target measure
- Nonlinear model reduction on metric spaces. Application to one-dimensional conservative PDEs in Wasserstein spaces
- Statistical shape analysis of brain arterial networks (BAN)
- Statistical data analysis in the Wasserstein space
- Depth and outliers for samples of sets and random sets distributions
- Minimax estimation of smooth optimal transport maps
- The transport map computed by iterated function system
- Central limit theorems for entropy-regularized optimal transport on finite spaces and statistical applications
- Estimation of Wasserstein distances in the spiked transport model
- Regularized optimal transport and the rot mover's distance
- A Variational Model Dedicated to Joint Segmentation, Registration, and Atlas Generation for Shape Analysis
- Geodesic PCA in the Wasserstein space by convex PCA
- Machine learning and optimal transport: some statistical and algorithmic tools
- Independent Nonlinear Component Analysis
Uses Software
This page was built for publication: Geodesic PCA versus Log-PCA of Histograms in the Wasserstein Space
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4610140)
