Independent component analysis for tensor-valued data
From MaRDI portal
Abstract: In preprocessing tensor-valued data, e.g. images and videos, a common procedure is to vectorize the observations and subject the resulting vectors to one of the many methods used for independent component analysis (ICA). However, the tensor structure of the original data is lost in the vectorization and, as a more suitable alternative, we propose the matrix- and tensor fourth order blind identification (MFOBI and TFOBI). In these tensorial extensions of the classic fourth order blind identification (FOBI) we assume a Kronecker structure for the mixing and perform FOBI simultaneously on each direction of the observed tensors. We discuss the theory and assumptions behind MFOBI and TFOBI and provide two different algorithms and related estimates of the unmixing matrices along with their asymptotic properties. Finally, simulations are used to compare the method's performance with that of classical FOBI for vectorized data and we end with a real data clustering example.
Recommendations
Cites work
- scientific article; zbMATH DE number 45789 (Why is no real title available?)
- scientific article; zbMATH DE number 1391247 (Why is no real title available?)
- A Multilinear Singular Value Decomposition
- A survey of multilinear subspace learning for tensor data
- Advanced Linear Algebra
- Characteristics of multivariate distributions and the invariant coordinate system
- Deflation-Based FastICA With Adaptive Choices of Nonlinearities
- Dimension folding PCA and PFC for matrix-valued predictors
- Eigenvectors of a kurtosis matrix as interesting directions to reveal cluster structure
- Existence and uniqueness of the maximum likelihood estimator for models with a Kronecker product covariance structure
- Fourth moments and independent component analysis
- Geodesic Convexity and Covariance Estimation
- Independent component analysis for tensor-valued data
- Invariant co-ordinate selection (with discussion)
- Maximum likelihood estimation for the tensor normal distribution: Algorithm, minimum sample size, and empirical bias and dispersion
- Models with a Kronecker product covariance structure: estimation and testing
- On Estimation of Covariance Matrices With Kronecker Product Structure
- On dimension folding of matrix- or array-valued statistical objects
- On invariant coordinate system (ICS) functionals
- On multilinear principal component analysis of order-two tensors
- Regularized Matrix Regression
- Robust Kronecker Product PCA for Spatio-Temporal Covariance Estimation
- Sliced Inverse Regression for Dimension Reduction
- Squared skewness minus kurtosis bounded by 186/125 for unimodal distributions
- Statistical Analysis of Financial Data in S-Plus
- Structured Lasso for regression with matrix covariates
- Tensor Decompositions and Applications
- Tensor Regression with Applications in Neuroimaging Data Analysis
- Tensor sliced inverse regression
- Tests for Kronecker envelope models in multilinear principal components analysis
- The multilinear normal distribution: introduction and some basic properties
Cited in
(10)- Independent component analysis for tensor-valued data
- Independent component analysis for multivariate functional data
- Fast tensorial JADE
- Tensor Canonical Correlation Analysis With Convergence and Statistical Guarantees
- On the usage of joint diagonalization in multivariate statistics
- Projected estimation for large-dimensional matrix factor models
- Application of the sequential matrix diagonalization algorithm to high-dimensional functional MRI data
- Independent component analysis: a statistical perspective
- JADE for Tensor-Valued Observations
- scientific article; zbMATH DE number 6906972 (Why is no real title available?)
This page was built for publication: Independent component analysis for tensor-valued data
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1679572)
