Two-stage estimation and simultaneous confidence band in partially nonlinear additive model (Q2051519)

The authors focus on the estimation and inference in partially nonlinear additive models. By integrating spline approximation and local smoothing, they propose a two-stage estimating approach in which the profile nonlinear least square method is used to estimate parameters and additive functions. Under some regularity conditions they establish the asymptotic normality of the parametric estimators and achieve an optimal nonparametric convergence rate of the fitted functions. To make inference on the nonparametric functions from the whole, the authors furthermore construct simultaneous confidence bands, and further propose an empirical bootstrap-based confidence band for the practical implementation. Finally the performance of the proposed method is illustrated in both a simulation study and real world applications.