Multilinear discriminant analysis using tensor-tensorproducts

DOI10.22124/JMM.2023.22841.2026MaRDI QIDQ6156172zbMATH OpenOpenAlexFDO

Authors F. Dufrenois, A. El Ichi, Khalide Jbilou

Publication date 13 June 2023

Full work available at URL https://arxiv.org/abs/2203.00967

zbMATH Keywords

Krylov subspaceslinear tensor equationstensor L-product

Mathematics Subject Classification ID

Multilinear algebra, tensor calculus (15A69) Ill-posedness and regularization problems in numerical linear algebra (65F22) Iterative numerical methods for linear systems (65F10)

Abstract: Multilinear Discriminant Analysis (MDA) is a powerful dimension reduction method specifically formulated to deal with tensor data. Precisely, the goal of MDA is to find mode-specific projections that optimally separate tensor data from different classes. However, to solve this task, standard MDA methods use alternating optimization heuristics involving the computation of a succession of tensor-matrix products. Such approaches are most of the time difficult to solve and not natural, highligthing the difficulty to formulate this problem in fully tensor form. In this paper, we propose to solve multilinear discriminant analysis (MDA) by using the concept of transform domain (TD) recently proposed in cite{Kilmer2011}. We show here that moving MDA to this specific transform domain make its resolution easier and more natural. More precisely, each frontal face of the transformed tensor is processed independently to build a separate optimization sub-problems easier to solve. Next, the obtained solutions are converted into projective tensors by inverse transform. By considering a large number of experiments, we show the effectiveness of our approach with respect to existing MDA methods.

Recommendations

• MPCA and MDA via Einstein product
• A survey of multilinear subspace learning for tensor data
• Large Margin Low Rank Tensor Analysis
• On multilinear principal component analysis of order-two tensors
• Extracting the optimal dimensionality for local tensor discriminant analysis

Cites work

• scientific article; zbMATH DE number 7404606 (Why is no real title available?)
• A Fast Algorithm for Deblurring Models with Neumann Boundary Conditions
• Discrete cosine transform LSQR methods for multidimensional ill-posed problems
• Facial recognition using tensor-tensor decompositions
• Factorization strategies for third-order tensors
• Gait recognition using linear time normalization
• Global Golub-Kahan bidiagonalization applied to large discrete ill-posed problems
• Kernel Fisher Discriminants for Outlier Detection
• One class proximal support vector machines
• Solving sparse non-negative tensor equations: algorithms and applications
• Tensor Decompositions and Applications
• Tensor-tensor products with invertible linear transforms
• The trace ratio optimization problem
• Third-order tensors as operators on matrices: a theoretical and computational framework with applications in imaging
• Trace optimization and eigenproblems in dimension reduction methods.

Cited in

(5)

• MPCA and MDA via Einstein product
• Using tensor products to detect unconditional label dependence in multilabel classifications
• Kronecker-decomposable robust probabilistic tensor discriminant analysis
• Spectral computation with third-order tensors using the t-product
• On numerical range of tensor via c-product

This page was built for publication: Multilinear discriminant analysis using tensor-tensorproducts

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6156172)