B1967
Title: Bootstrap inference in functional linear regression models with scalar response
Authors: Xiongtao Dai - University of California, Berkeley (United States) [presenting]
Hyemin Yeon - Kent State University (United States)
Daniel Nordman - Iowa State University (United States)
Abstract: A fundamental problem of bootstrap inference is investigated for functional linear regression models (FLRMs). The work further develops the central limit theorem (CLT) and residual bootstrap in FLRMs by generalizing, refining, and correcting existing results. The theoretical results on bootstrap and the CLT are novel and apply either conditionally or unconditionally on data regressors. We also develop a new method for obtaining simultaneous inference based on residual bootstrap and provide theoretical supports. Numerical studies verify our theory, showing that the bootstrap performs better than normal approximations, and also suggest a rule of thumb for setting the truncation levels. The bootstrap method is illustrated with an application to wheat spectrum data.
