半参数空间变系数回归模型的Back-Fitting估计

Back-Fitting Procedure for Semiparametric Spatially Varying-Coefficient Regression Model

针对半参数空间变系数回归模型给出了一种估计方法-后向拟合估计,该方法可得到模型中常值系数估计量的精确解析表达式,广泛的数值模拟表明所提出的估计方法对估计常值系数具有满意的精度和稳定性,最后,利用该方法分析了一个实际的例子.

This paper proposes a novel procedure for fitting the semiparametric spatially varying-coefficient regression model, by which an explicit expression for estimators of the constant coefficients in the model can be obtained. Extensive simulations are then conducted to examine the performance of the proposed fitting procedure and the results demonstrate that the estimators for the constant coefficients are quite accurate and stable, finally, we analysis a real data.