摘要
利用加权主成分分析法、GIS的趋势分析工具和Morans’I指数分析2009年黄河下游沿岸109个县域经济发展的空间差异,并运用MATLAB7.0的SOFM神经网络模型进行经济发展水平分类。结果表明:黄河下游沿岸县域经济发展的空间分异显著,宏观上呈现出西南—东北方向的U型趋势和东南—西北方向的倒U型趋势;县域经济空间集聚特征明显,东营、济南各县域以及荥阳为经济核心区,而豫东、菏泽等中原经济区东部平原县为经济塌陷区。109个县域的经济发展水平可以分为5类,其空间分布总体上符合圈层结构理论。用主成分得分作为输入,参考层次聚类结果确定神经元数是SOFM网络取得良好分类效果的前提。
The spatial differences of economic development of 109 counties along the Lower Yellow River in 2009 have been analyzed by weighted principal component analysis,trend analysis tool and Moran's I index of GIS.On the basis,the classification of level of economic development is realized by the SOFM neural network model in MATLAB 7.0.Study shows that,firstly,there are significant differences for the economic development of counties along the Lower Yellow River.In the global space,it is the U-trend in the southwest-northeast and the inverted U-trend in the southeast-northwest,and the spatial concentration of the county economy is very significant as was indicated by Moran scatter plot.Moreover,Xingyang and counties in Dongying and Jinan are economically core area,while counties in Yudong Plain and Heze,in the eastern Zhongyuan economic zone,are economically "collapse" area.Finally,the level of economic development of 109 counties can be divided into five categories with the principal component scores as the input of the SOFM network,which is generally in line with the circle-type spatial structure theory.At the same time,it is a prerequisite to achieve good classification results by referring the results of hierarchical clustering to determine the number of neurons in the SOFM network.
出处
《经济地理》
CSSCI
北大核心
2012年第3期16-21,共6页
Economic Geography
基金
国家教育部人文社会科学重点研究基地重大项目(10JJDZONGHE015)
国家自然科学基金面上项目(41171438)
省部共建河南大学研究项目(SBGJ090111)
关键词
县域经济发展
空间分异
SOFM分类
空间自相关分析
黄河下游沿岸区
economic strength of counties
spatial differentiation
SOFM clustering technique
spatial autocorrelation analysis
region along the Lower Yellow River