Abstract
Formulations of multidimensional scaling (MDS) in dually flat spaces are proposed. First the space supposed in the classical MDS is extended to a tangent space around a specific point in a dually flat space. We see that Riemannian metric of the tangent point plays a key role in the extension. Next, in order to remove the restriction of symmetry in dissimilarities, the affine connection is incorporated. We pay attention to the fact that it is an affine connection term that causes an asymmetry in dissimilarities in infinitesimal space. To mitigate the difficulty in treating the affine connection term, an approximation is shown and we can see the effect of the affine connection term to modify the effective Riemannian metric. Finally a numerical example is shown.
Similar content being viewed by others
References
Borg, I., Groenen, P.: Modern Multidimensional Scaling, 2nd edn. Springer, New York (2005)
Gower, J.C.: Some distance properties of latent root and vector methods used in multivariate analysis. Biometrika 53, 325–338 (1966)
Torgerson, W.S.: Theory and Methods of Scaling. Wiley, New York (1958)
Chino, N.: A graphical technique for representing the asymmetric relationships between \(N\) objects. Behaviormetrika 5, 23–40 (1978)
Chino, N.: A generalized inner product model for the analysis of assymetry. Behaviormetrika 27, 25–46 (1990)
Chino, N., Shiraiwa, K.: Geometrical structures of some non-distance models for asymmetric MDS. Behaviormetrika 20, 35–47 (1993)
Chino, N.: A brief survey of asymmetric MDS and some open problems. Behaviormetrika 39, 127–165 (2012)
Escoufier, Y., Grorud, A.: Analyse factorielle des matrices carées non-symétriques. In: Diday, E., et al. (eds.) Data Analysis and Informatics, pp. 263–276. North-Holland, Amsterdam (1980)
Gower, J.C.: The analysis of asymmetry and orthogonality. In: Barra, J.R., Brodeau, F., Romer, G., van Cutsem, B. (eds.) Recent Developments in Statistics, pp. 109–123. North-Holland, Amsterdam (1977)
Kiers, H.A.L., Takane, Y.: A generalization of GIPSCAL of nonsymmetric data. J. Classif. 11, 79–99 (1994)
Krumhansl, C.L.: Concerning the applicability of geometric models to similarity data: the interrelationship between similarity and spatial density. Psychol. Rev. 85, 445–463 (1987)
Okada, A., Imaizumi, T.: Nonmetric multidimensional scaling of asymmetric proximities. Behaviormetrika 21, 81–96 (1987)
Saito, T., Takeda, S.: Multidimensional scaling of asymmetric proximity: model and method. Behaviormetrika 28, 49–80 (1990)
Saito, T.: Analysis of asymmetric proximity matrix by a model of distance and additive terms. Behaviormetrika 29, 45–60 (1991)
Saito, T., Yadohisa, H.: Data Analysis of Asymmetric Structures. Marcel Dekker, New York (2005)
Tversky, A.: Features of similarity. Psychol. Rev. 84, 327–352 (1977)
Mucherino, A., Lavor, C., Liberti, L., Maculan, N. (eds.): Distance Geometry: Theory, Methods, and Applications. Springer, New York (2013)
Weinberger, K.Q., Saul, L.K.: Unsupervised learning of image manifolds by semidefinite programming. Int. J. Comput. Vis. 70, 77–90 (2006)
Lee, J.A., Verleysen, M.: Nonlinear Dimensionality Reduction. Springer, New York (2010)
Tenenbaum, J.B., Silva, V., Langford, J.C.: A global geometric framework for nonlinear dimensionality reduction. Science 290, 2319–2323 (2000)
Roweis, S.T., Saul, L.K.: Nonlinear dimensionality reduction by locally linear embedding. Science 290, 2323–2326 (2000)
Schölkopf, B., Smola, A., Müller, K.R.: Nonlinear component analysis as a kernel eigenvalue problem. Neural Comput. 10, 1299–1319 (1998)
Kumagai, A.: Analysis of asymmetric relational data based on classical multidimensional scaling (in Japanese). Trans. Jpn Soc. Ind. Appl. Math. 20, 57–66 (2010)
Kumagai, A.: Handling asymmetry of relational data from a viewpoint of classical multidimensional scaling (in Japanese). Trans. Jpn Soc. Ind. Appl. Math. 21, 165–174 (2011)
Kumagai, A.: Extension of classical MDS to treat dissimilarities not satisfying axioms of distance. Jpn. J. Indust. Appl. Math. 31, 111–124 (2014)
Plasencia-Calaña, Y., et al.: On the informativeness of asymmetric dissimilarities. Lect. Notes Comput. Sci. 7923, 75–89 (2013)
Amari, S.: Methods of Information Geometry. American Mathematical Society, Providence (2001)
Cichocki, A., Amari, S.: Families of alpha- beta- and gamma-divergences: flexible and robust measures of similarities. Entropy 12, 1532–1568 (2010)
Bregman, L.M.: The relaxation method of finding the common point of convex sets and its application to the solution of problems in convex programming. USSR Comput. Math. Math. Phys. 7, 200–217 (1967)
Csiszár, I.: Eine informationstheoretische Ungleichung und ihre Anwendung auf den Beweis der Ergodizitat von Markoffschen Ketten, Magyar. Tud. Akad. Mat. Kutato Int. Kozl 8, 85–108 (1963)
Zhang, J.: Dual scaling between comparison and reference stimuli in multidimensional psychological space. J. Math. Psychol. 48, 409–424 (2004)
Banerjee, A., Merugu, S., Dhillon, I.S., Ghosh, J.: Clustering with Bregman divergences. J. Mach. Learn. Res. 6, 1705–1749 (2005)
Amari, S.: Differential geometry derived from divergence functions: information geometry approach. In: Deza, M., Petitjean, M., Markov K. (eds.) Abstracts of the International Conference “Mathematics of distances and applications”, Varna, Bulgaria, 2–5 July 2012. http://www.foibg.com/ibs_isc/ibs-25/ibs-25-f.pdf
Indow, T.: An approach to geometry of visual space with no a priori mapping functions: multidimensional mapping according to Riemannian metrics. J. Math. Psychol. 26, 204–236 (1982)
Indow, T.: A critical review of Luneburg’s model with regard to global structure of visual space. Psychol. Rev. 96, 430–453 (1991)
Cheng, X., Shen, Z., Zhou, Y.: On a class of locally dually flat Finsler metrics. Int. J. Math. 21, 1531–1543 (2010)
Sato, Y.: An analysis of sociometric data by MDS. In: Matusita, K. (ed.) Statistical Theory and Data Analysis II, pp. 385–396. North-Holland, Amsterdam (1988)
Brachat, J., Comon, P., Elias, T.: Symmetric tensor decomposition. Linear Algebra Appl. 433, 1851–1872 (2010)
Comon, P., Golub, G., Lim, L.-H., Mourrain, B.: Symmetric tensors and symmetric tensor rank. SIAM J. Matrix Anal. Appl. 30, 1254–1279 (2008)
Acknowledgments
The author greatly appreciates the valuable comments of the reviewers.
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix A: Additional examples of numerical calculation
In this appendix, additional dissimilarity data are numerically analyzed according to Sect. 6.
First the following circulant dissimilarity matrix is analyzed. Table 2 shows the result.
Next the following non-circulant dissimilarity matrix is analyzed. Table 3 shows the result.
Appendix B: An example of graphical representation
In addition to the decomposition of Riemannian metric
through common eigenvalue decomposition, let us consider a symmetric tensor decomposition [38, 39]
where \( \xi _1 \ge \cdots \ge \xi _r > 0\), \(\sum _{l=1}^r (w_l^i)^2 = 1\). Then the Eq. (30) is written as
where \(x_{\iota l} = \sqrt{\mu _l} u^i_l \eta _{i\iota }, \ y_{\iota l} = \root 3 \of {\xi _l} w^i_l \eta _{i\iota }\).
The first term in the Eq. (43), the symmetric part of \(d_{\iota \kappa }\), corresponds to the squared distance in classical MDS. The calculation of coordinates \(x_{\iota l}\) and their graphical representation are the same as in classical MDS.
The remaining is the treatment of the second term in the Eq. (43), the antisymmetric part of \(d_{\iota \kappa }\). In addition to \(x_{\iota l}\), \(y_{\iota l}\) stands for the coordinate of the object \(\iota \) in the \(l\)-th dimension. Here, as an example, a two-dimensional representation is presented. Supposing that the coordinate of the object \(n\) be zero, the antisymmetric part of \(d_{\iota n}\) is written for each \(\iota \) as follows:
This is rewritten for graphical representation as
where \(s_\iota = y_{\iota 1} + y_{\iota 2}, \ t_\iota = \sqrt{3} (y_{\iota 1} - y_{\iota 2})\). Thus we can represent the antisymmetric parts by plotting \((s_\iota , \ t_\iota )\) in a plane. For each object \(\iota \), the antisymmetric part of \(d_{\iota n}\) is represented as a quarter of the product of the value of \(s_\iota \) and the squared distance between \(\iota \) and \(n\).
About this article
Cite this article
Kumagai, A. Multidimensional scaling in dually flat spaces. Japan J. Indust. Appl. Math. 32, 51–63 (2015). https://doi.org/10.1007/s13160-014-0164-2
Received:
Revised:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s13160-014-0164-2