CANDECOMP/PARAFAC Decomposition of High-order Tensors Through Tensor Reshaping
Authors:
- Anh Huy PhanAPAnh Huy Phan
- Petr TichavskyPTPetr Tichavsky
- Andrzej CichockiACAndrzej Cichocki
Preprint
Abstract
In general, algorithms for order-3 CANDECOMP/-PARAFAC (CP), also coined canonical polyadic decomposition (CPD), are easily to implement and can be extended to higher order CPD. Unfortunately, the algorithms become computationally demanding, and they are often not applicable to higher order and relatively large scale tensors. In this paper, by exploiting the uniqueness of CPD and the relation of a tensor in Kruskal form and its unfolded tensor, we propose a fast approach to deal with this problem. Instead of directly factorizing the high order data tensor, the method decomposes an unfolded tensor with lower order, e.g., order-3 tensor. On basis of the order-3 estimated tensor, a structured Kruskal tensor of the same dimension as the data tensor is then generated, and decomposed to find the final solution using fast algorithms for the structured CPD. In addition, strategies to unfold tensors are suggested and practically verified in the paper.Keywords
All Related Versions
This publication has 28 references indexed in Scilit:
- Some New Results on Orthogonally Constrained CandecompJournal of Classification, 2011
- Fast truncation of mode ranks for bilinear tensor operationsNumerical Linear Algebra with Applications, 2011
- Recent advances on the use of separated representationsInternational Journal for Numerical Methods in Engineering, 2009
- Tensor decompositions, alternating least squares and other talesJournal of Chemometrics, 2009
- A Kronecker product approximate preconditioner for SANsNumerical Linear Algebra with Applications, 2004
- Application of the three‐way decomposition for matrix compressionNumerical Linear Algebra with Applications, 2002
- Slowly converging parafac sequences: Swamps and two‐factor degeneraciesJournal of Chemometrics, 1994
- Analysis of Individual Differences in Multidimensional Scaling Via an N-way Generalization of “Eckart-Young” DecompositionPsychometrika, 1970
- Some Mathematical Notes on Three-Mode Factor AnalysisPsychometrika, 1966
- Multiple Invariants and Generalized Rank of a P‐Way Matrix or TensorJournal of Mathematics and Physics, 1928