B0446
Title: A unified approach for hypothesis testing in functional linear models
Authors: Yinan Lin - National University of Singapore (Singapore)
Zhenhua Lin - National University of Singapore (Singapore) [presenting]
Abstract: A unified approach is developed for hypothesis testing in various types of functional linear models, such as scalar-on-function, function-on-function, function-on-scalar models, that have a wide range of applications in functional data analysis. In addition, the proposed test can handle models of mixed types, such as models with both functional and scalar/vector predictors. Unlike most existing methods that rest on large-sample distributions of test statistics, the proposed method leverages the technique of bootstrapping max statistics and exploits the variance decay property that is an inherent feature of functional data to achieve superior numerical performance, especially when the sample size is limited. Theoretical guarantees on the validity and consistency of the proposed test are provided uniformly for a class of test statistics.