ICMAT DataLab Seminar Series

"Reproducing kernel methods in statistics: some recent contributions" 

by Antonio Cuevas, Departamento de Matemáticas (UAM) and ICMAT.

Date and Time: Friday, 24 November 2023, 11am (CET)

Place: Sala Roja IFT, C/ Nicolás Cabrera, nº 13-15. Campus de Cantoblanco, UAM, Madrid.

Abstract: The notion of Reproducing Kernel Hilbert Space (RKHS) is well-known in the mathematical community since the middle of last century. In fact, RKHS's are a very useful tool in different fields, ranging from mathematical analysis and approximation theory to functional statistics. In this expository talk I will review some recent contributions of our research group on the use of RKHS in statistical methods for functional data. The considered topics include supervised classification, regression (linear and logistic), definition of suitable statistical distances (of Mahalanobis type), and two-sample tests for functional data. 

The results reviewed in the talk are included in joint works with J. R. Berrendero, B. Bueno-Larraz, J. Cárcamo, A. Cholaquidis, A. Coín, L.A. Rodríguez and J.L. Torrecilla.