Nonparametric Methods for the Homogenization of Functional Data with Applications in Public Health and Environment

M. Oviedo de la Fuente, R. Fernández-Casal, M. A. Florez

In numerous problems in public health, environment, and energy, information is recorded continuously, generating data with a functional structure, such as daily pollution curves, intraday electricity demand profiles, or real-time IoT signals. These trajectories often exhibit outliers, missing observations, and temporal dependence, which complicates the direct application of Functional Data Analysis (FDA) techniques.
An applied review of nonparametric strategies for the homogenization and analysis of functional data is presented, combining smoothing under dependence, functional kriging for the reconstruction of incomplete curves, and detection of outlier curves using depth measures. The methodologies are illustrated with climatological, energy, and public health data. The aim is to provide a practical and unified framework for the robust treatment of functional data in real-world applications.

Keywords: Nonparametric methods, functional data, public health, energy demand, outliers, missing data

SI Statistics and the Sustainable Development Goals (SDGs): Methodological Approaches and Applications in Health, Climate Action, and Energy
September 3, 2026  9:00 AM
Aula 30

M. Díaz-Gay

C. García-Martos, A. M. Alonso, E. Martínez, S. Monge, R. Rodríguez Fernández, M. J. Sierra, J. Sánchez de la Mata

S. J. Koopman, C. Amoroso, C. García-Martos

M. Wiper, C. Ausín, A. Sarhadi, M. Pérez García