【作者】 吴浩；

【导师】 李清泉； 花向红；

【作者基本信息】 武汉大学 ， 大地测量学与测量工程， 2003， 硕士

【摘要】 本文围绕多元统计理论在GIS非空间特性分析的应用来进行研究,主要探讨了利用因子分析和聚类分析两种方法来实现对非空间特性的建模过程。首先讨论当前多元统计理论与GIS结合的必要性,提出了两者的三种结合模式。详细阐述采用因子分析对多维复杂数据群进行结构简化和消除变量间相关性的分析建模过程,包括因子分析适宜性评价、因子求解方法对比、因子个数确定原则、因子旋转作用与方法以及因子得分的计算等。随后探讨通过聚类分析对多维复杂数据群实现分类,以系统聚类法和动态聚类法为重点,详细论述各自建模的具体过程和算法以及聚类个数的评定标准。 在此基础之上,采用因子分析和聚类分析相结合,对影响和决定县区国民经济可持续发展多项指标之间的相互关系进行定量分析。以广东省统计局发布的99年各县区国民经济年报数据为例进行分析,找出影响广东省各县区经济可持续发展的主要因子及其构成;然后以因子分析的结果——因子得分作为聚类分析的初始数据源,对广东省99年各县区经济可持续发展状况进行分类。为制定适合国民经济可持续发展的政策和评估各县区国民经济可持续发展情况等工作提供理论依据和技术支持。更多还原

【Abstract】 Non-spatial information analysis, which is used to solve those problems related to spatial analysis, is an analysis method for Non-spatial characteristics of spatial data and an important component of spatial analysis in GIS. However, in the GIS alteration course from project-driven to data-driven, non-spatial information analysis nowadays puts up lots of shortcomings, such as short of powerful analysis ability and ignoring those characteristics of data actual nature, which are far from away to provide assistant decision-making support. So, it is necessary to import multivariate statistical theory to Non-spatial information analysis. In this paper, factor analysis and cluster analysis are used to improve non-spatial information analysis.Firstly, this paper discusses the combination difficulties of multivariate statistical theory and GIS. At the same time, three kinds of combination models are provided and compared. Secondly, factor analysis are made use to reduce structure and eliminate relativity for mass data. The course of factor analysis are discussed, including the evaluation of analysis suitability, the factor computation ways, the numbers of factor, the factor rotation methods, the calculation of factor score and so on. Afterwards, cluster analysis is used to classify mass data. The paper are attached most importance to both Hierarchical cluster and fast cluster. The course of modeling and arithmetic are discussed in detail. Especially, those four assessing criterions such as RSQ, SPRSQ, PSF and PST2, are provided and compared. Based on factor analysis and cluster analysis, an analysis model of economy sustainable development assessment is established and made on Guangdong Province as an example. After every county economy data in 1999 are analysed, several factors of economy sustainable development are gained and those counties are classified by the above result of factor analysis. It can bring forward lots of reasonable suggestions to make policies of suitable economy sustainable development and assess its development condition.更多还原