Title: Variable selection in varying-coefficient functional linear models
Authors: Hidetoshi Matsui - Shiga University (Japan) [presenting]
Abstract: Varying-coefficient functional linear models consider the relationship between a scalar response and functional predictors, where the coefficient functions depend on an exogeneous variable. It then accounts for the relation of the predictors and the response varying with the exogeneous variable. We consider the problem of variable selection in the varying-coefficient functional linear model with multiple functional predictors. To solve this problem, we apply the group lasso-type regularization that induces sparsity. The proposed method is applied to the analysis of agricultural data. In particular, we select environmental factors that relate to the crop yield of multi-stage tomatoes.