Uniqueness proof for a family of models sharing features of Tucker's three-mode factor analysis and PARAFAC/candecomp

作者: Richard A. Harshman , Margaret E. Lundy

DOI: 10.1007/BF02296963

关键词: MathematicsGeneralityMode (statistics)Principal component analysisMathematical optimizationApplied mathematicsMultidimensional scalingUniquenessSymmetry (geometry)Orientation (vector space)Rotation (mathematics)

摘要: Some existing three-way factor analysis and MDS models incorporate Cattell's “Principle of Parallel Proportional Profiles”. These can—with appropriate data—empirically determine a unique best fitting axis orientation without the need for separate rotation stage, but they have not been general enough to deal with what Tucker has called “interactions” among dimensions. This article presents proof considerably more parallel profiles model which incorporates interacting The model, Xk=AADk HBDk B', does assume symmetry in data or interactions factors. A second is presented symmetrically weighted case (i.e., whereADk=BDk). generality these allows one impose successive restrictions obtain several useful special cases, including PARAFAC2 DEDICOM.

参考文章(28)
M. E. Lundy, J. B. Kruskal, R. A. Harshman, How 3-MFA data can cause degenerate parafac solutions, among other relationships Multiway data analysis. pp. 115- 122 ,(1989)
J. B. Kruskal, Rank, decomposition, and uniqueness for 3-way and n -way arrays Multiway data analysis. pp. 7- 18 ,(1989)
Ledyard R Tucker, Some mathematical notes on three-mode factor analysis Psychometrika. ,vol. 31, pp. 279- 311 ,(1966) , 10.1007/BF02289464
Richard A. Harshman, Margaret E. Lundy, PARAFAC: parallel factor analysis Computational Statistics & Data Analysis. ,vol. 18, pp. 39- 72 ,(1994) , 10.1016/0167-9473(94)90132-5
Michael R. W. Dawson, Richard A. Harshman, The multidimensional analysis of asymmetries in alphabetic confusion matrices: Evidence for global-to-local and local-to-global processing Perception & Psychophysics. ,vol. 40, pp. 370- 383 ,(1986) , 10.3758/BF03208196
Ledyard R Tucker, Relations Between Multidimensional Scaling and Three-Mode Factor Analysis. Psychometrika. ,vol. 37, pp. 3- 27 ,(1972) , 10.1007/BF02291410
Jos M. F. ten Berge, Henk A. L. Kiers, Some uniqueness results for PARAFAC2 Psychometrika. ,vol. 61, pp. 123- 132 ,(1996) , 10.1007/BF02296962
Pieter M. Kroonenberg, Jan de Leeuw, Principal component analysis of three-mode data by means of alternating least squares algorithms Psychometrika. ,vol. 45, pp. 69- 97 ,(1980) , 10.1007/BF02293599
William Meredith, Notes on factorial invariance Psychometrika. ,vol. 29, pp. 177- 185 ,(1964) , 10.1007/BF02289699