ML Estimation of Spatial Panel Data Geographically Weighted Regression Model

摘要

In this article, the data base of Geographically Weighted Regression (shorted for GWR) model was expanded to panel data. Based on local bandwidth theory, spatial panel data GWR model Maximum Likelihood (ML) estimation , which is under the specification of pooled effect,fixed effects, and random effects, was also deduced in this article, as solved the problem that GWR model can only be used with the cross-sectional data and covered the flaw of global bandwidth insufficiency, so that the estimation could go better with economic actuality.