An Analysis on Spatial Differentiation of County-level Economy in Shenyang Metropolitan Area Based on BP Neural Network
利用BP神经网络和ESDA技术，分析了2005和2010年沈阳经济区县域经济的空间分异，发现南北差异缩小、相似经济水平县域空间分散，区位和政策是主要影响因素。
On the basis of the BP neural network and ESDA technology,this paper uses 9 economic indicators to analyses the spatial differentiation of county-level economy in the Shenyang Metropolitan Area in 2005 and 2010.The results are as follows.Firstly,there is a narrowing tendency of the spatial differentiation of county-level economy between the southern and northern Liaoning province in 2005 and 2010.The economic evaluation values of some industrialized counties such as Haicheng and Dashiqiao are high from 2005 to 2010,the values of the central-south counties such as Dengta and Liaoyang decline,and the northeast counties such as Kaiyuan and Diaobingshan increase at the same time.Secondly,the decline of Moran’s I from 2005 to 2010 shows that counties of similar economic development levels are spatially scattered,while counties of better economic levels distribute along the main transportation axes.Thirdly,in terms of local differences,in 2005,Xiuyan and Huanren belong toH-Htype,and Zhangwu and Kangping areL-L type counties.In 2010,Kaiyuan isH-Ltype,while the other counties don’t show the extremely spatial distribution characteristics,which means that the counties of different economic levels are of a random spatial distribution pattern.At last,the paper holds the view point that the location condition is the basic factor for the formation of the spatial differentiation of county-level economy,the top-down regional development policies perform as external impetus,and the polarization and diffusion influence of regional central cities are the important force.