An Efficient Approach for Estimating Parameters and Nonparametric Functions in Spatio-temporal Semi-parametric Regression Models

This work investigates statistical inference for both the mean and covariance functions in semiparametric models for complex spacetime-dependent data. A new kernel estimator for the spatiotemporally correlated data is proposed to estimate nonparametric functions, and we show the new method can improve estimation bias and efficiency of nonparametric functions from existing kernel methods such as the local linear regression (LLR).

A simulation result

Figure 1: Pointwise Variance and Mean Square Error of the estimations for g(u). (1) LLR (Liu et al., 2021, JMVA) and the proposed method.