Functional single-index model for partially observed functional data

Silvia Novo

Universitad Carlos III de Madrid

Thursday 9th January 2025

13:00:00

Aula 3.1.3

Link to seminar will become active ten minutes before due starting time: Click HERE to refresh page

In the context of regression with functional data, one of the most popular semiparametric models is the functional single-index model. For this model, which involves a scalar response and a single functional covariate, Novo, Aneiros, and Vieu (2019, Journal of Nonparametric Statistics) proposed an automatic estimation procedure capable of adapting to the local characteristics of the data. However, the studied methodology remains restrictive as it assumes the functional covariate is fully observed. In this ongoing work, we compare different reconstruction methods for the functional covariate, such as those developed by Kneip and Liebl (2020, The Annals of Statistics) and Palummo et al. (2024, Environmental and Ecological Statistics), and evaluate their impact on model estimation through simulations and applications to real-world datasets.