SuperMap GIS地理加权回归模型研究与实现

地理加权回归分析是对普通线性回归模型的扩展,将空间数据的地理位置嵌入线性回归参数之中,以此来研究空间关系的空间异质性或空间非平稳性,属于局部空间分析模型.通过地理加权回归分析可以确定两种或两种以上变量间相互依赖的定量关系,局部区域的参数估计可以得到地理空间存在的不同空间关系,核函数的选取规则和带宽参数的验证方法也是本文研究的内容.

Research and Implementation of the Geographically Weighted Regression Model with SuperMap

Geographically weighted regression analysis is the extension of ordinary linear regression model. The spatial data geograph-ical position is embedded into the linear regression parameters, to study the spatial relations of spatial heterogeneity and spatial no sta-tionary, which belongs to the local spatial analysis model. By means of geographically weighted regression analysis, the quantitative relationship between two or more variables can be determined, and the parameters of local regions are estimated with different spatial relationships. The selection rules of kernel function and the verification method of determining bandwidth parameters are also studied.