Pspatreg: R Package for Semiparametric Spatial Autoregressive Models Articles uri icon

publication date

  • November 2024

start page

  • 2598

issue

  • 22

volume

  • 12

International Standard Serial Number (ISSN)

  • 2227-7390

abstract

  • This article introduces the R package pspatreg, which is publicly available for download from the Comprehensive R Archive Network, for estimating semiparametric spatial econometric penalized spline (P-Spline) models. These models can incorporate a nonparametric spatiotemporal trend, a spatial lag of the dependent variable, independent variables, noise, and time-series autoregressive noise. The primary functions in this package cover the estimation of P-Spline spatial econometric models using either Restricted Maximum Likelihood (REML) or Maximum Likelihood (ML) methods, as well as the computation of marginal impacts for both parametric and nonparametric terms. Additionally, the package offers methods for the graphical display of estimated nonlinear functions and spatial or spatiotemporal trend maps. Applications to cross-sectional and panel spatial data are provided to illustrate the package"s functionality.

subjects

  • Statistics

keywords

  • semiparametric regression; spatial econometrics; p-splines