A FURTHER EXPLORATION INTO THE ROBUSTNESS OF SPATIAL AUTOCORRELATION SPECIFICATIONS*
通过不同预测框架的蒙特卡洛实验，检验了地质统计模型优于权重矩阵模型这一结论的稳健性，并给出质疑该结论的实验结果。
ABSTRACT In a recent study, the robustness of linear models with various spatial autocorrelation specifications was assessed through Monte Carlo experiments, and the geostatistical models were concluded to dominate the weight matrix models in the prediction. The present study tests the soundness of this conclusion with a different framework for prediction and presents some experimental results that can call into doubt the dominance of the geostatistical models over the weight matrix models.