Non-Gaussian component analysis: testing the dimension of the signal subspace

DOI10.1007/978-3-030-48814-7_6zbMATH Open1455.62105arXiv2007.02751OpenAlexW3045604959WikidataQ109772885 ScholiaQ109772885MaRDI QIDQ5141231FDOQ5141231

Klaus Nordhausen, Una Radojičić

Publication date: 18 December 2020

Published in: Analytical Methods in Statistics (Search for Journal in Brave)

Abstract: Dimension reduction is a common strategy in multivariate data analysis which seeks a subspace which contains all interesting features needed for the subsequent analysis. Non-Gaussian component analysis attempts for this purpose to divide the data into a non-Gaussian part, the signal, and a Gaussian part, the noise. We will show that the simultaneous use of two scatter functionals can be used for this purpose and suggest a bootstrap test to test the dimension of the non-Gaussian subspace. Sequential application of the test can then for example be used to estimate the signal dimension.

Full work available at URL: https://arxiv.org/abs/2007.02751

Recommendations

• Independent Component Analysis and Blind Signal Separation
• In search of non-Gaussian components of a high-dimensional distribution
• Asymptotic and bootstrap tests for subspace dimension
• Independent Component Analysis and Blind Signal Separation
• Non-Gaussian component analysis using entropy methods

zbMATH Keywords

bootstrapindependent component analysisscatter matrixorder determinationsignal subspace

Mathematics Subject Classification ID

Factor analysis and principal components; correspondence analysis (62H25) Estimation in multivariate analysis (62H12) Signal theory (characterization, reconstruction, filtering, etc.) (94A12) Signal detection and filtering (aspects of stochastic processes) (60G35)

Cites Work

• Invariant co-ordinate selection (with discussion)
• ICS for Multivariate Outlier Detection with Application to Quality Control
• Principal component analysis.
• Robust Estimation of a Location Parameter
• What is Projection Pursuit?
• Optimization and testing in linear non‐Gaussian component analysis
• Linear Non-Gaussian Component Analysis Via Maximum Likelihood
• Projection pursuit
• Redescending \(M\)-estimates of multivariate location and scatter
• \(M\)-functionals of multivariate scatter
• New algorithms for \(M\)-estimation of multivariate scatter and location
• A cautionary note on robust covariance plug-in methods
• Fourth moments and independent component analysis
• Title not available (Why is that?)
• A new algorithm of non-Gaussian component analysis with radial kernel functions
• Symmetrised M-estimators of multivariate scatter
• Multivariate Models and the First Four Moments
• On the Computation of Symmetrized M-Estimators of Scatter
• Uniqueness of Non-Gaussianity-Based Dimension Reduction

Cited In (4)

• Asymptotic and bootstrap tests for subspace dimension
• On the usage of joint diagonalization in multivariate statistics
• Numerical Considerations and a new implementation for invariant coordinate selection
• Colored Subspace Analysis

Uses Software

• R
• ICS
• SpatialNP
• JADE
• BSSasymp
• extraDistr
• png
• REPPlab
• ICSOutlier
• ICtest
• RcppRoll

This page was built for publication: Non-Gaussian component analysis: testing the dimension of the signal subspace

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5141231)