A0901
Title: A projection-based diagnostic test for generalized functional regression models
Authors: Hua Liang - George Washington University (United States) [presenting]
Abstract: A novel diagnostic test is proposed to check goodness-of-fit for generalized functional regression models. The proposed test is free of any distribution assumptions and can be used for various classical functional regression models. However, it is based on independence in distribution and hence includes mean-based and higher-order moment-based tests as special cases. The proposed test avoids any subjective selection of tuning parameters by integrating over the directions along which the functional variables project. As a result, it enhances the local power and overcomes the infinite dimensionality problem simultaneously. A rather simple implementation procedure is developed. The performance of the proposed test is evaluated through theory and extensive simulation studies. The proposed procedure is also applied to analyze Canadian Weather data and Chinese Air Pollution data, resulting in several interesting models which achieve higher interpretability and estimation accuracy than the existing methods.
