Artículo
Abstract:
Canonical correlation analysis (CCA) is a dimension-reduction technique in which two random vectors from high dimensional spaces are reduced to a new pair of low dimensional vectors after applying linear transformations to each of them, retaining as much information as possible. The components of the transformed vectors are called canonical variables. One seeks linear combinations of the original vectors maximizing the correlation subject to the constraint that they are to be uncorrelated with the previous canonical variables within each vector. By these means one actually gets two transformed random vectors of lower dimension whose expected square distance has been minimized subject to have uncorrelated components of unit variance within each vector. Since the closeness between the two transformed vectors is evaluated through a highly sensitive measure to outlying observations as the mean square loss, the linear transformations we are seeking are also affected. In this paper we use a robust univariate dispersion measure (like an M-scale) based on the distance of the transformed vectors to derive robust S-estimators for canonical vectors and correlations. An iterative algorithm is performed by exploiting the existence of efficient algorithms for S-estimation in the context of Principal Component Analysis. Some convergence properties are analyzed for the iterative algorithm. A simulation study is conducted to compare the new procedure with some other robust competitors available in the literature, showing a remarkable performance. We also prove that the proposal is Fisher consistent. © 2014 Elsevier Inc.
Registro:
| Documento: | Artículo |
| Título: | A robust predictive approach for canonical correlation analysis |
| Autor: | Adrover, J.G.; Donato, S.M. |
| Filiación: | FAMAF, Universidad Nacional de Córdoba, CIEM and CONICET, Argentina Instituto de Cálculo, Universidad de Buenos Aires and CONICET, Argentina |
| Palabras clave: | Canonical correlation analysis; M-scales; Mean relative prediction error; S-estimation |
| Año: | 2015 |
| Volumen: | 133 |
| Página de inicio: | 356 |
| Página de fin: | 376 |
| DOI: | http://dx.doi.org/10.1016/j.jmva.2014.09.007 |
| Título revista: | Journal of Multivariate Analysis |
| Título revista abreviado: | J. Multivariate Anal. |
| ISSN: | 0047259X |
| CODEN: | JMVAA |
| Registro: | https://bibliotecadigital.exactas.uba.ar/collection/paper/document/paper_0047259X_v133_n_p356_Adrover |
Referencias:
- Berrendero, J.R., Contribuciones a la Teoría de la Robustez respecto al Sesgo, , (Ph.D. thesis), Universidad Carlos III de Madrid (in Spanish)
- Boente, G., Orellana, L., A robust approach to common principal components (2001) Trends in Mathematics, pp. 117-145. , Birkhäuser Verlag
- Branco, J.A., Croux, C., Filzmoser, P., Oliveira, M.R., Robust canonical correlations: a comparative study (2005) Comput. Statist., 20 (2), pp. 203-229
- Croux, C., Dehon, C., Estimators of the multiple correlation coeficient: Local robustness and confidence intervals (2003) Statist. Papers, 44, pp. 315-334
- Das, S., Sen, P.K., Canonical correlations (1998) Encyclopedia of Biostatistics, pp. 468-482. , John Wiley & Sons, Ltd., New York, P. Armitage, T. Colton (Eds.)
- Davies, P.L., Asymptotic behavior of S-estimators of multivariate location estimators and dispersion matrices (1987) Ann. Statist., 15, pp. 1269-1292
- Filzmoser, P., Dehon, C., Croux, C., Outlier resistant estimators for canonical correlation analysis (2000) COMPSTAT: Proceedings in Computational Statistics, pp. 301-306. , Physica-Verlag, Heldelber, J.G. Betlehem, P.G.M. van der Heijden (Eds.)
- Furrer, R., Genton, M.G., Aggregation-cokriging for highly multivariate spatial data (2011) Biometrika, 98, pp. 615-631
- Hoffman, K., Kunze, R., (2009) Linear Algebra, , PHI Learning
- Hotelling, H., Relations between two sets of variables (1936) Biometrika, 28, pp. 321-377
- Karnel, G., Robust canonical correlation and correspondence analysis (1991) The Frontiers of Statistical Scientific and Industrial Applications, (Vol. II of the Proceedings of ICOSCO-I, The First International Conference on Statistical Computing), pp. 335-354. , American Sciences Press, Strasbourg
- Maronna, R.A., Principal components and orthogonal regression based on robust scales (2005) Technometrics, 47 (3), pp. 264-273
- Rao, C., Toutenberg, H., (1999) Linear Models: Least Squares and Alternatives, , Springer
- Romanazzi, M., Influence in canonical correlation analysis (1992) Psychometrika, 57, pp. 237-259
- Rousseeuw, P.J., Multivariate estimation with high breakdown point (1985) Mathematical Statistics and Applications, Vol. B, pp. 283-297. , Reidel, Dordrecht, W. Grossmann, G. Pflug, I. Vincze, W. Wertz (Eds.)
- Seber, G.A.F., (1984) Multivariate Observations, , John Wiley & Sons, Inc
- Taskinen, S., Croux, C., Kankainen, A., Ollila, E., Oja, H., Influence functions and efficiencies of the canonical correlation and vector estimates based on scatter and shape matrices (2006) J. Multivariate Anal., 97, pp. 359-384
- ten Berge, J.M.F., On the equivalence of two oblique congruence rotation methods and orthogonal approximations (1979) Psychometrika, 44, pp. 359-364
- Wold, H., Nonlinear estimation by iterative least squares procedures (1966) A Festschrift for J. Neyman, pp. 411-444. , Wiley, New York, F.N. Davis (Ed.)
- Yohai, V.J., García Ben, M., Canonical variables as optimal predictors (1980) Ann. Statist., 8 (4), pp. 865-869
- Zamar, R.H., Bias robust estimation in orthogonal regression (1992) Ann. Statist., 20 (4), pp. 1875-1888
Citas:
---------- APA ----------
Adrover, J.G. & Donato, S.M. (2015) . A robust predictive approach for canonical correlation analysis. Journal of Multivariate Analysis, 133, 356-376.http://dx.doi.org/10.1016/j.jmva.2014.09.007
---------- CHICAGO ----------
Adrover, J.G., Donato, S.M. "A robust predictive approach for canonical correlation analysis" . Journal of Multivariate Analysis 133 (2015) : 356-376.http://dx.doi.org/10.1016/j.jmva.2014.09.007
---------- MLA ----------
Adrover, J.G., Donato, S.M. "A robust predictive approach for canonical correlation analysis" . Journal of Multivariate Analysis, vol. 133, 2015, pp. 356-376.http://dx.doi.org/10.1016/j.jmva.2014.09.007
---------- VANCOUVER ----------
Adrover, J.G., Donato, S.M. A robust predictive approach for canonical correlation analysis. J. Multivariate Anal. 2015;133:356-376.http://dx.doi.org/10.1016/j.jmva.2014.09.007
