文档介绍：兰州大学硕士研究生学位论文基于 ESDA 的甘肃省县域经济空间差异分析
摘要
区域经济差异一直是地理学和经济学研究的热点问题之一,过去的研究多以
省份和城市作为研究对象,很少涉及到较小的经济单元,例如县域和乡镇。为了
揭示甘肃省在微观层次的经济差异并考虑到数据的可获取性,本文选取县域作为
研究分析的基本空间单元。
本文选择甘肃省县域作为研究对象,选取 8 个指标并利用主成份分析方法和
人工神经网络模型对县域经济发展水平进行评价,同时借助 GIS 软件分析甘肃省
县域经济发展水平的区域差异。而后,基于 ESDA 全局和局部空间自相关分析,
利用甘肃省 1992~2006 年县域人均 GDP 数据,对甘肃省县域总体和局部空间差
异的变化趋势、特征与成因进行了初步探索。结果表明甘肃省经济空间自相关特
性明显,经济发展水平相似的地区在空间上集聚分布,呈现出“哑铃”形的空间
格局,即经济发达地区集中分布在河西地区,经济发展水平较低的地区集中分布
在陇中、陇南、陇东地区。
最后针对分析结果提出缩小县域经济发展水平的措施和建议。
关键词:人工神经网络;县域经济;区域经济差异;ESDA;空间自相关;甘肃
省
I
兰州大学硕士研究生学位论文基于 ESDA 的甘肃省县域经济空间差异分析
ABSTRACT
Regional economic disparity is always the hot spot problem of geography and
economics. The researches in the past often take provinces and cities as research
object and hardly refer to smaller spatial cell, such as counties and villages. In order to
reveal economic disparities in microcosmic aspect of Gansu province and considering
the gain of data, the article take counties as basic spatial cell of research..
The evaluation of economic development level of counties in Gansu was
researched with the method of artificial works and ponent
analysis, adopting 8 economic indexes, meanwhile, GIS techniques were applied to
reveal the differentiations and the strategy of the development of counties.
Based on the global and local spatial autocorrelation analyses of Exploratory
Spatial Data Analysis (ESDA),this paper investigates the spatial dynamics of regional
disparities at the county level in Gansu by analyzing per capita GDP
data,1992- results show that spatial autocorrelation of Gansu Province
at the county level is significant. That is ,the relatively high (low) developed county
tends to be located nearby other high (low) counties more often than expected due to
random chance .The relatively high developed counties are located at Hexi corridor
and the relatively low ones are located at central, eastern and southern Gansu, which
presents a “dumbbell” pattern.
At last ,