Central limit theorems for classical multidimensional scaling

DOI10.1214/20-EJS1720zbMATH Open1445.62117arXiv1804.00631MaRDI QIDQ2192306FDOQ2192306

Minh Tang, Gongkai Li, Nicolas Charon, Carey E. Priebe

Publication date: 17 August 2020

Published in: Electronic Journal of Statistics (Search for Journal in Brave)

Abstract: Classical multidimensional scaling is a widely used method in dimensionality reduction and manifold learning. The method takes in a dissimilarity matrix and outputs a low-dimensional configuration matrix based on a spectral decomposition. In this paper, we present three noise models and analyze the resulting configuration matrices, or embeddings. In particular, we show that under each of the three noise models the resulting embedding gives rise to a central limit theorem. We also provide compelling simulations and real data illustrations of these central limit theorems. This perturbation analysis represents a significant advancement over previous results regarding classical multidimensional scaling behavior under randomness.

Full work available at URL: https://arxiv.org/abs/1804.00631

zbMATH Keywords

central limit theoremperturbation analysisdissimilarity matrixclassical multidimensional scaling

Mathematics Subject Classification ID

Statistical aspects of information-theoretic topics (62B10) Classification and discrimination; cluster analysis (statistical aspects) (62H30) Estimation in multivariate analysis (62H12) Central limit and other weak theorems (60F05)

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?)
• Diffusion maps
• Modern multidimensional scaling. Theory and applications.
• The Rotation of Eigenvectors by a Perturbation. III
• Automatic dimensionality selection from the scree plot via the use of profile likelihood
• Matrix estimation by universal singular value thresholding
• Analysis of individual differences in multidimensional scaling via an \(n\)-way generalization of ``Eckart-Young decomposition
• High-Dimensional Probability
• Local Multidimensional Scaling for Nonlinear Dimension Reduction, Graph Drawing, and Proximity Analysis
• A useful variant of the Davis–Kahan theorem for statisticians
• Multidimensional scaling. I: Theory and method
• A class of invariant consistent tests for multivariate normality
• Measures of multivariate skewness and kurtosis with applications
• Multidimensional scaling.
• Energy statistics: a class of statistics based on distances
• User-friendly tail bounds for sums of random matrices
• The Dissimilarity Representation for Pattern Recognition
• Laplacian Eigenmaps for Dimensionality Reduction and Data Representation
• A limit theorem for scaled eigenvectors of random dot product graphs
• Large deformation diffeomorphic metric curve mapping
• The Euclidian Distance Matrix Completion Problem
• Solving Euclidean distance matrix completion problems via semidefinite progrmming
• A remark on global positioning from local distances
• Localization from incomplete noisy distance measurements
• Single subject incomplete designs for nonmetric multidimensional scaling
• Distance Shrinkage and Euclidean Embedding via Regularized Kernel Estimation
• Exact Reconstruction of Euclidean Distance Geometry Problem Using Low-Rank Matrix Completion

Cited In (2)

• The out-of-sample problem for classical multidimensional scaling
• Singular vector distribution of sample covariance matrices

Uses Software

• energy
• mdsr

This page was built for publication: Central limit theorems for classical multidimensional scaling

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2192306)