基于GWR方法的山东省县域空间经济数据分析

Based on the economic data, Shandong 140 counties and districts were analyzed to study whether these economic data with a spatial correlation. In this paper we build two models which are ordinary least squares and geographic weighted regression to analyze the data of Shandong county economy. GDP not only reflects the economic level of a country, but also reflect the comprehensive national strength and wealth of a country.So it is often regarded as a measure of the country's economic situation and is regarded as the best indicator. In this paper, the per capita GDP is regarded as the dependent variable.We use the global auto-correlation and ordinary least squares to make sure that there is no redundancy and have spatial correlation of the independent variables. After repeated operations, finalized the the independent variables are as follows: industry structure variable (the ratio of the secondary industry and tertiary industry ), extroversion of economic variables (per capita actual utilization of foreign capital), capital input variables (per capita investment in fixed assets), human resource variable (the proportion of employees in the total population). In this paper,we make the economic data have a global autocorrelation analysis,ordinary least squares analysis,geographic weighted regression analysis.Studies have shown that: One of them is that the ratio of the secondary industry and tertiary industry and per capita GDP tend to be inversely related in coastal areas, while the inland areas of the ratio of the secondary industry and tertiary industry is proportional to the GDP per capita; the per capita actual utilization of foreign capital is proportional to per capita GDP, the per capita actual utilization of foreign capital impact on the inland areas is more significant than that in coastal areas.Per capita investment in fixed assets is proportional to the per capita GDP; proportion of employees is proportional to the per capita GDP.