Geographically weighted regression estimation of the linear response and plateau function
2022
Lambert, Dayton M. | Cho, Whoi
The objective of this paper is to modify the linear geographically weighted regression (GWR) estimator to accommodate the discontinuous join point, or “knot”, of the linear response with plateau (LRP). This is the first application to estimate a LRP site-specific crop response function (SSCRF) model with GWR. The data used in this heuristic application are from a variable rate nitrogen (VRN) trial for corn. Results from a partial budget comparison of uniform and VRN management are sensitive with respect to the choice of kernel used to weight yield observations. Challenges remain with respect to the availability and cost of equipment required to implement precision fertilizer recommendations based on spatially varying SSCRF. Ex post spatial cluster analysis of site-specific fertilizer prescriptions may be one approach for simplifying complex application maps into larger management units.
显示更多 [+] 显示较少 [-]