|
Abstract:
|
In regression analysis, spline surfaces can be used to capture spatial dependence in a spatial linear regression model, without imposing a parametric covariance structure. However, including a spline component may impact the accuracy and precision of the estimated regression coefficients. The resulting bias is analogous to the bias seen with spatially correlated random effects in the scenario of spatial confounding. In this talk, we investigate such impacts in spline-based semiparametric regression for spatial data. We discuss estimators' behavior and propose a correction based on the spline smoothing parameter. Numerical examples, including a case study in precision agriculture, will also be presented.
|
