CFE 2019: Start Registration
View Submission - CMStatistics
Title: Varying-coefficient functional additive models Authors:  Hidetoshi Matsui - Shiga University (Japan) [presenting]
Abstract: Varying-coefficient functional linear models consider the relationship between a scalar response and a functional predictor, where the coefficient function depends on another variable. It then accounts for the influence of the variable to the response. We extend the varying-coefficient functional linear model to the framework of an additive model and propose a varying-coefficient functional additive model. It represents the nonlinear relationship between the response and the predictor by introducing nonlinear functions of functional predictor. We estimate the varying-coefficient functional additive model by the penalized likelihood method. The effectiveness of the proposed model is investigated through simulation and real data analysis.