TA Links
TA stands for tensor approximation and it is a mathematical framework, which is used in various areas for compact data representation. Here a couple of helpful links on the basic definitions, models and algorithms are given.
Books, Reviews and Surveys
- Kolda & Bader, 2009: Tensor Decompositions and Applications
- De Lathauwer, 2009: A Survey of Tensor Methods
- Kroonenberg, 2008: Applied Multiway Data Analysis
- Smilde et al., 2004: Multi-way Analysis: Applications in the Chemical Sciences
- Moravitz, 2004: Tensor decompositions workshop discussion notes
Theory and Background
Notation
- Kolda & Bader, 2009: Tensor Decompositions and Applications
- Smilde et al., 2004: Multi-way Analysis: Applications in the Chemical Sciences
- Kiers, 2000: Towards a standardized notation and terminol- ogy in multiway analysis
- De Lathauwer et al, 2000b: On the best rank-1 and rank-(R1,R2,...,RN) approximation of higher-order tensors
- De Lathauwer et al, 2000a: A multilinear singular value decomposition
Models
- Tucker model: Tucker, 1963: Implications of Factor Analysis of Three-way Matrices for Measurement of Change, Tucker, 1964: The Extension of Factor Analysis to Three-dimensional Matrices, and Tucker, 1966: Some Mathematical Notes on Three-mode Factor Analysis
- CP model: Caroll & Chang, 1970: Analysis of Individual Differences in Multidimensional Scaling via an N-way Generalization of Eckart--Young Decompositions and Harshman, 1970: Foundations of the PARAFAC Procedure: Models and Conditions for an Explanatory Multi--modal Factor Analysis
- Block-based TA: De Lathauwer, 2008: Decompositions of a Higher-Order Tensor in Block Terms---Part II: Definitions and Uniqueness
- Others see Kolda & Bader, 2009: Tensor Decompositions and Applications
Algorithms
Alternating least-squares (ALS) algorithms:
- Tucker1 (three-mode SVD): Tucker, 1966: Some Mathematical Notes on Three-mode Factor Analysis
- Three-mode PCA ALS: Kroonenberg & De Leeuw, 1980: Principal Component Analysis of Three-mode Data by Means of Alternating Least Squares Algorithms and Kroonenberg, 1983: Three-mode principal component analysis: Theory and applications
- TUCKALS: Ten Berge et al., 1987: Some Additional Results on Principal Components Analysis of Three-mode Data by Means of Alternating Least Squares Algorithms
- Higher-order singular value decomposition (HOSVD): De Lathauwer et al, 2000a: A multilinear singular value decomposition
- Higher-order orthogonal iteration (HOOI) and higher-order power method (HOPM): De Lathauwer et al, 2000b: On the best rank-1 and rank-(R1,R2,...,RN) approximation of higher-order tensors
- Block-based TA ALS: De Lathauwer & Nion, 2008: Decompositions of a Higher-Order Tensor in Block Terms---Part III: Alternating Least Squares Algorithms
- For improvements on the above algorithms see literature in Kolda & Bader, 2009: Tensor Decompositions and Applications
Libraries
- Tensor classes in vmmlib
- Tensor Toolbox for MATLAB
- Tensorlab
- Others see Kolda & Bader, 2009: Tensor Decompositions and Applications