Advertisement

Psychometrika

, Volume 61, Issue 1, pp 133–154 | Cite as

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

  • Richard A. Harshman
  • Margaret E. Lundy
Article

Abstract

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

Key words

Parallel proportional profiles intrinsic axes DEDICOM PARAFAC2 Cattell trilinear models quadrilinear models factor rotation problem multidimensional scaling principal components oblique confactor 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Carroll, J. D., & Chang, J. J. (1970). Analysis of individual differences in multidimensional scaling via an N-way generalization of “Eckart-Young” decomposition.Psychometrika, 35, 283–319.Google Scholar
  2. Carroll, J. D., & Wish, M. (1974). Models and methods for three-way multidimensional scaling. In D. H. Krantz, R. C. Atkinson, R. D. Luce, & P. Suppes (Eds.),Contemporary developments in mathematical psychology: Vol. 2 (pp. 57–105). San Francisco: W. H. Freeman.Google Scholar
  3. Cattell, R. B. (1944). “Parallel Proportional Profiles” and other principles for determining the choice of factors by rotation.Psychometrika, 9, 267–283.Google Scholar
  4. Cattell, R. B. (1978).The scientific use of factor analysis in behavioral and life sciences. New York: Plenum.Google Scholar
  5. Cattell, R. B., & Cattell, A. K. S. (1955). Factor rotation for proportional profiles: Analytical solution and an example.British Journal of Statistical Psychology, 8, 83–92.Google Scholar
  6. Dawson, M. R. W., & Harshman, R. A. (1986). The multidimensional analysis of asymmetries in alphabetic confusion matrices: Evidence for global-to-local and local-to-global processing.Perception & Psychophysics, 40, 370–383.Google Scholar
  7. de Leeuw, J., & Pruzansky, S. (1978). A new computational method to fit the weighted Euclidean distance model.Psychometrika, 43, 479–490.CrossRefGoogle Scholar
  8. Dunn, T. R., & Harshman, R. A. (1982). A multidimensional scaling model for the size-weight illusion.Psychometrika, 47, 25–45.CrossRefGoogle Scholar
  9. Harshman, R. A. (1970). Foundations of the PARAFAC procedure: Models and conditions for an “explanatory” multi-modal factor analysis.UCLA Working Papers in Phonetics, 16, 1–84. (University Microfilms No. 10,085)Google Scholar
  10. Harshman, R. A. (1972a). Determination and proof of minimum uniqueness conditions for PARAFAC1.UCLA Working Papers in Phonetics, 22, 111–117. (University Microfilms No. 10,085)Google Scholar
  11. Harshman, R. A. (1972b). PARAFAC2: Mathematical and technical notes.UCLA Working Papers in Phonetics, 22, 30–44. (University Microfilms No. 10,085)Google Scholar
  12. Harshman, R. A. (1992).Unusual uniqueness properties of a three-way factor analysis model that incorporates “higher order” factors. Unpublished manuscript.Google Scholar
  13. Harshman, R. A., & DeSarbo, W. (1984). An application of PARAFAC to a small sample problem, demonstrating preprocessing, orthogonality constraints, and split-half diagnostic techniques. In H. G. Law, C. W. Snyder, Jr., J. A. Hattie, & R. P. McDonald (Eds.),Research methods for multimode data analysis (pp. 602–642). New York: Praeger.Google Scholar
  14. Harshman, R. A., & Lundy, M. E. (1984). The PARAFAC model. In H. G. Law, C. W. Snyder, Jr., J. A. Hattie, & R. P. McDonald (Eds.),Research methods for multimode data analysis (pp. 122–215). New York: Praeger.Google Scholar
  15. Harshman, R. A., & Lundy, M. E. (1994). PARAFAC: Parallel factor analysis.Computational Statistics & Data Analysis, 18, 39–72.CrossRefGoogle Scholar
  16. Harshman, R. A., Green, P. E., Wind, Y., & Lundy, M. E. (1982). A model for the analysis of asymmetric data in marketing research.Marketing Science, 1, 205–242.Google Scholar
  17. Kroonenberg, P. M. (1983).Three mode principal component analysis. Leiden: DSWO Press.Google Scholar
  18. Kroonenberg, P. M., & de Leeuw, J. (1980). Principal components analysis of three-mode data by means of alternating least-squares algorithms.Psychometrika, 5, 69–97.Google Scholar
  19. Kruskal, J. B. (1976). More factors than subjects, tests than treatments: An indeterminacy theorem for canonical decomposition and individual differences scaling.Psychometrika, 41, 281–293.CrossRefGoogle Scholar
  20. Kruskal, J. B. (1977). Three-way arrays: Rank and uniqueness of trilinear decompositions, with application to arithmetic complexity and statistics.Linear Algebra and Its Applications, 18, 95–138.CrossRefGoogle Scholar
  21. Kruskal, J. B. (1981). Multilinear models for data analysis.Behaviormetrika, 10, 1–20.Google Scholar
  22. Kruskal, J. B. (1989). Rank, decomposition, and uniqueness for 3-way and N-way arrays. In R. Coppi & S. Bolasco (Eds.),Multiway data analysis (pp. 7–18). Amsterdam: North Holland.Google Scholar
  23. Kruskal, J. B., Harshman, R. A., & Lundy, M. E. (1989). How 3-MFA data can cause degenerate PARAFAC solutions, among other relationships. In R. Coppi & S. Bolasco (Eds.),Multiway data analysis (pp. 115–130). Amsterdam: North Holland.Google Scholar
  24. Leurgans, S. E. (1991).Multilinear models: Array formulas (Technical Report 473). Columbus, OH: Ohio State University, Department of Statistics.Google Scholar
  25. Leurgans, S. E., Ross, R. T., & Abel, R. B. (1993). A decomposition for 3-way arrays.SIAM Journal on Matrix Analysis and Applications, 14, 1064–1083.CrossRefGoogle Scholar
  26. Meredith, W. (1964). Notes on factorial invariance.Psychometrika, 29, 177–185.Google Scholar
  27. Mitchell, B. C., & Burdick, D. S. (1994). Slowly converging PARAFAC sequences: Swamps and two-factor degeneracies.Journal of Chemometrics, 8, 155–168.CrossRefGoogle Scholar
  28. ten Berge, J. M. F., & Kiers, H. A. L. (1996). Some uniqueness results for PARAFAC2.Psychometrika, 61, 123–132.CrossRefGoogle Scholar
  29. Tucker, L. R. (1963). Implications of factor analysis of three-way matrices for measurement of change. In C. W. Harris (Ed.),Problems in measuring change (pp. 122–137). Madison, Wisconsin: University of Wisconsin Press.Google Scholar
  30. Tucker, L. R. (1964). The extension of factor analysis to three-dimensional matrices. In N. Frederiksen & H. Gulliksen (Eds.),Contributions to mathematical psychology (pp. 109–127). New York: Holt, Rinehart, & Winston.Google Scholar
  31. Tucker, L. R. (1966). Some mathematical notes on three-mode factor analysis.Psychometrika, 31, 279–311.PubMedGoogle Scholar
  32. Tucker, L. R. (1972). Relations between multidimensional scaling and three-mode factor analysis.Psychometrika, 37, 3–27.CrossRefGoogle Scholar

Copyright information

© The Psychometric Society 1996

Authors and Affiliations

  • Richard A. Harshman
    • 1
  • Margaret E. Lundy
    • 1
  1. 1.Psychology DepartmentUniversity of Western OntarioLondonCanada

Personalised recommendations