Statistical inference in functional semiparametric spatial autoregressive model

• Main Authors: Gaosheng Liu, Yang Bai
• Format: Article
• Language: English
• Published: AIMS Press 2021-07-01
• Series: AIMS Mathematics
• Subjects: semiparametric spatial autoregressive model; functional data analysis; b-spline approximation; generalized method of moments
• Online Access: https://www.aimspress.com/article/doi/10.3934/math.2021633?viewType=HTML

Semiparametric spatial autoregressive model has drawn great attention since it allows mutual dependence in spatial form and nonlinear effects of covariates. However, with development of scientific technology, there exist functional covariates with high dimensions and frequencies containing rich information. Based on high-dimensional covariates, we propose an interesting and novel functional semiparametric spatial autoregressive model. We use B-spline basis function to approximate the slope function and nonparametric function and propose generalized method of moments to estimate parameters. Under certain regularity conditions, the asymptotic properties of the proposed estimators are obtained. The estimators are computationally convenient with closed-form expression. For slope function and nonparametric function estimators, we propose the residual-based approach to derive its pointwise confidence interval. Simulation studies show that the proposed method performs well.