A weak degeneracy revealing decomposition for the CANDECOMP/PARAFAC model

Journal of Chemometrics

Volumn 24, Issue 2, 2010, Pages 57-66

  Rocci, Roberto ^a   Giordani, Paolo ^b  

^a UNIVERSITY OF ROME TOR VERGATA   (Italy)

^b SAPIENZA UNIVERSITY OF ROME   (Italy)

Author keywords

Border rank; CANDECOMP PARAFAC; Degeneracy; Tucker3

Indexed keywords

BORDER RANK; CANDECOMP/PARAFAC; CANONICAL FORM; DATA ARRAY; DEGENERACY; LOSS FUNCTIONS; THREE WAY DATA; TUCKER3; TWO-COMPONENT;

EID: 77049124762     PISSN: 08869383     EISSN: 1099128X     Source Type: Journal    
DOI: 10.1002/cem.1272     Document Type: Article

References (42)

1. Carroll JD, Chang JJ. Analysis of individual differences in multidimensional scaling via an n-way generalization of Eckart-Young decomposition. Psychometrika 1970; 35: 283-319.
2. Harshman RA. Foundations of the Parafac procedure: models and conditions for an 'explanatory' multimodal factor analysis. UCLA Work. Pap. Phon. 1970; 16: 1-84.
3. Kruskal JB. Three-way arrays: rank and uniqueness of trilinear decompositions, with applications to arithmetic complexity and statistics. Linear Algebra Appl. 1977; 18: 95-138.
4. ten Berge JMF, Smilde AK. Non-triviality and identification of a constrained Tucker3 analysis. J. Chemom. 2002; 16: 609-612.
5. Harshman RA, Ladefoged P, Goldstein L. Factor analysis of tongue shapes. J. Acoust. Soc. Am. 1977; 62: 693-707.
6. Krijnen WP, Ten Berge JMF. A constrained PARAFAC method for positive manifold data. Appl. Psychol. Meas. 1992; 16: 295-305.
7. Harshman RA, DeSarbo WS. An application of PARAFAC to a small sample problem, demonstrating preprocessing, orthogonality constraints, and split-half diagnostic techniques. In Research Methods for Multi-mode Data Analysis, Law HG, Snyder CW Jr, Hattie J, McDonald RP (eds). Praeger: New York, 1984; 602-642.
8. Smilde A, Bro R, Geladi P. Multi-way Analysis: Applications in the Chemical Sciences. Wiley: 2004.
9. Andersen AH, Rayens WS. Structure-seeking multilinear methods for the analysis of fMRI data. NeuroImage 2004; 22: 728-739.
10. Krijnen WP, Dijkstra TK, Stegeman A. On the non-existence of optimal solutions and the occurrence of 'degeneracy' in the Candecomp/ Parafac model. Psychometrika 2008; 73: 431-439.
11. Harshman RA, Lundy ME. Data preprocessing and the extended PARAFAC model. In Research Methods for Multimode Data Analysis, Law HG, Snyder CW Jr, Hattie J, McDonald RP (eds). Praeger: New York, 1984; 216-284.
12. Tucker LR. Some mathematical notes on three-mode factor analysis. Psychometrika 1966; 31: 279-311.
13. Lundy ME, Harshman RA, Kruskal JB. A two stage procedure incorporating good features of both trilinear and quadrilinear models. In Multiway Data Analysis, Coppi R, Bolasco S (eds). Elsevier: Amsterdam, 1989; 123-130.
14. Kruskal JB, Rank decomposition, and uniqueness for 3-way and N-way arrays. In Multiway Data Analysis, Coppi R, Bolasco S (eds). Elsevier: Amsterdam, 1989; 7-18.
15. Tomasi G, Bro G. A comparison of algorithms for fitting the PARAFAC model. Comput. Stat. Data Anal. 2006; 50: 1700-1734.
16. De Lathauwer L. A link between the canonical decomposition in multilinear algebra and simultaneous matrix diagonalization. SIAM J. Matrix Anal. Appl. 2006; 28: 642-666.
17. Jiang T, Sidiropoulos ND. Kruskal's permutation lemma and the identification of Candecomp/Parafac and bilinear models with constant modulus constraints. IEEE Transactions on Signal Processing 2004; 52: 2625-2636.
18. Stegeman A, Ten Berge JMF, De Lathauwer L. Sufficient conditions for uniqueness in Candecomp/Parafac and Indscal with random component matrices. Psychometrika 2006; 71: 219-229.
19. De Silva V, Lim L-H. Tensor rank and the ill-posedness of the best low-rank approximation problem. SIAM J. Matrix Anal. Appl. 2008; 30: 1084-1127.
20. Kruskal JB, Harshman RA, Lundy ME. How 3-MFA data can cause degenerate PARAFAC solutions, among other relationships. In Multiway Data Analysis, Coppi R, Bolasco S (eds). Elsevier: Amsterdam, 1989; 115-122.
21. Mitchell BC, Burdick DS. Slowly converging Parafac sequences: swamps and two-factor degeneracies. J. Chemom. 1994; 8: 155-168.
22. Paatero P. Construction and analysis of degenerate Parafac models. J. Chemom. 2000; 14: 285-299.
23. Rayens WS, Mitchell BC. Two-factor degeneracies and a stabilization of Parafac. Chemom. Intell. Lab. Syst. 1997; 38: 173-181.
24. Stegeman A. Degeneracy in Candecomp/Parafac explained for p×p×2 arrays of rank p+1 or higher. Psychometrika 2006; 71: 483-501.
25. Stegeman A. Degeneracy in Candecomp/Parafac and Indscal explained for several three-sliced arrays with a two-valued typical rank. Psychometrika 2007; 72: 601-619.
26. Stegeman A. Low-rank approximation of generic p q 2 arrays and diverging components in the Candecomp/Parafac model. SIAM J. Matrix Anal. Appl. 2008; 30: 988-1007.
27. Stegeman A. Using the simultaneous generalized Schur decomposition as a Candecomp/Parafac algorithm for ill-conditioned data. J. Chemom. 2009; 23: 385-392.
28. Stegeman A, De Lathauver L. A method to avoid diverging components in the Candecomp/Parafac model for generic I×J×2 arrays. SIAM J. Matrix Anal. Appl. 2009; 30: 1614-1638.
29. ten Berge JMF, Kiers HAL, De Leeuw J., Explicit Candecomp/Parafac solutions for a contrived 2×2×2 array of rank three. Psychometrika 1988; 53: 579-584.
30. Bini D. Border rank of a p×q×2 tensor and the optimal approximation of a pair of bilinear forms, In Automata, Languages and Programming (Lecture Notes in Computer Science, 85), de Bakker JW, van Leeuwen J (eds). Springer: New York, 1980; 98-108.
31. Kroonenberg PM. Applied Multiway Data Analysis. Wiley Series in Probability and Statistics: Hoboken, NJ, 2008.
32. Paatero P. A weighted non-negative least squares algorithm for three-way'PARAFAC' factor analysis. Chemom. Intell. Lab. Syst. 1997; 38: 223-242.
33. De Lathauwer L, De Moor B, Vandewalle J. Computation of the canonical decomposition by means of a simultaneous generalized schur decomposition. SIAM J. Matrix Anal. Appl. 2004; 26: 295-327.
34. Kroonenberg PM, de Leeuw J. Principal component analysis of three-mode data by means of alternating least squares algorithms. Psychometrika 1980; 45: 69-97.
35. Rocci R. Three-mode factor analysis with binary core and orthonormality constraints. J. Ital. Stat. Soc. 1992; 3: 413-422.
36. ten Berge JMF. Kruskal's polynomial for 2×2×2 arrays and a generalization to 2×n×n arrays. Psychometrika, 1991;56: 631-636.
37. Bini D, Capovani M, Romani F, Lotti G. O(n2.7799) complexity for n×n approximate matrix multiplication. Inf. Process. Lett. 1979; 8: 234-235.
38. Bini D, Lotti G, Romani F. Approximate solutions for the bilinear form computational problem. SIAM J. Comput. 1980; 9: 692-697.
39. Bini D. Relations between exact and approximate bilinear algorithms. Applications. Calcolo 1980; 17: 87-97.
40. Gurden SP, Westerhuis JA, Bijlsma S, Smilde A. Modeling of spectroscopic batch process data using grey models to incorporate external information. J. Chemom. 2001; 15: 101-121.
41. Bro R. Multi-way analysis in the food industry. Models Algorithms and Applications, University of Amsterdam, 1998.
42. Bro R, Kiers HAL. A new efficient method for determining the number of components in PARAFAC models. J. Chemom. 2003; 17: 274-286.