A quantile based extension of functional principal component analysis

A. Mendez Civieta, J. Goldsmith

This work introduces an innovative framework for analyzing high-resolution physical activity data by combining functional data analysis with principal component techniques. The approach extends traditional methods by incorporating quantile regression to capture diverse distributional features, while a multilevel modeling structure accounts for variability both between and within individuals. This methodology offers a new avenue for robustly characterizing time-varying activity patterns in health-related research.

Keywords: Functional-data; quantile-regression; physical-activity; PCA

Big Data processing and analysis (TABiDa2)
June 10, 2025  5:10 PM
Sala 3. Maria Rúbies Garrofé

J. A. Salmerón Garrido

C. Yildirim, R. E. Lillo, A. M. Franco Pereira

P. González Barquero, R. E. Lillo Rodríguez, Á. Méndez Civieta