Testing Linearity in the Single Functional Index Model for Dependent Data
摘要
Motivated by checking the linearity assumption in the regression model with functional covariate and real response in the case of dependent data, a new test for the single functional index model is introduced. The critical region of the test is derived by a bootstrap procedure. The finite sample performances of the test are evaluated by a simulation study, and an application in the framework of functional time series forecasting is performed.