Simulation studies and breakdown analysis suggest that the proposed methods may perform better than the classical estimator under non-normal models and in the presence of outlying observations.
📌 3 Key Takeaways
🎯
Objective & Hypothesis:In this paper, the canonical correlation analysis for multivariate functional data is considered.
📊
Key Findings & Data:The analysis is based on the basis functions representation of the data.
💡
Conclusion & Impact:Simulation studies and breakdown analysis suggest that the proposed methods may perform better than the classical estimator under non-normal models and in the presence of outlying observations.
🔬Design:Modeling & Simulation
📍Scope:Comprehensive Sample
⚙️Variables:Basis functions representation · Canonical correlation · Functional data analysis
Abstract
In this paper, the canonical correlation analysis for multivariate functional data is considered. The analysis is based on the basis functions representation of the data. The use of non-orthogonal bases is available in contrast to the approach given in the literature. The robust estimation methods of the covariance matrix are also studied in the multivariate functional canonical correlation analysis. Simulation studies and breakdown analysis suggest that the proposed methods may perform better than the classical estimator under non-normal models and in the presence of outlying observations.
Aneiros, G., Bongiorno, E.G., Cao, R. and Vieu, P. (Eds) Functional statistics and related fields. Contributions to Statistics (Springer, 2017).
Bobelyn, E., Serban, A.S., Nicu, M., Lammertyn, J., Nicolai, B.M. and Saeys, W. Postharvest quality of apple predicted by NIR-spectroscopy: Study of the effect of biological variability on spectra and model performance. Postharvest Biol. Technol. 55, 133-143, 2010.
Boente, G., Barrera, M.S. and Tyler, D.E. A characterization of elliptical distributions and some optimality properties of principal components for functional data. J. Multivariate Anal. 131, 254-264, 2014.
Bongiorno, E.G., Salinelli, E., Goia, A. and Vieu, P. Contributions in infinite-dimensional statistics and related topics (Societa Editrice Esculapio, 2014).
Branco, J.A., Croux, C., Filzmoser, P. and Oliveira, M.R. Robust canonical correlations: A comparative study. Comput. Statist. 20, 203-229, 2005.
Chen, D., Hall, P. and Müller, H.G. Single and multiple index functional regression models with nonparametric link. Ann. Statist. 39, 1720-1747, 2011.
Collazos, J.A.A., Dias, R. and Zambom, A.Z. Consistent variable selection for functional regression models. J. Multivariate Anal. 146, 63-71, 2016.
Croux, C. and Haesbroeck, G. Influence function and efficiency of the minimum covariance determinant scatter matrix estimator. J. Multivariate Anal. 71, 161-190, 1999.
Davies, P.L. Asymptotic behavior of S-estimators of multivariate location parameters and dispersion matrices. Ann. Statist. 15, 1269-1292, 1987.
Davies, P.L. An efficient Fréchet-differentiable high breakdown multivariate location and dispersion estimator. J. Multivariate Anal. 40, 311-327, 1992.
Delaigle, A. and Hall, P. Achieving near perfect classification for functional data. J. R. Stat. Soc. Ser. B Stat. Methodol. 74, 267-286, 2012.
Febrero-Bande, M., Galeano, P. and González-Manteiga, W. Outlier detection in functional data by depth measures, with application to identify nbnormal NOx levels. Environmetrics 19, 331-345, 2008.
Ferraty, F. and Vieu, P. Nonparametric functional data analysis: Theory and practice (Springer, New York, 2006).
Fremdt, S., Horváth, L., Kokoszka, P. and Steinebach, J.G. Functional data analysis with increasing number of projections. J. Multivariate Anal. 124, 313-332, 2014. Górecki, T., Krzyśko, M., Waszak, Ł. and Wołyński, W. Selected statistical methods of data analysis for multivariate functional data. Statist. Papers 59, 153-182, 2018. Górecki, T. and Smaga, Ł.