空间数据统计分析的思想起源与应用演化  

The origin and application history of statistical analysis of spatial data

作　　者：赵永[1] ZHAO Yong(College of Environment and Planning,Henan University,Kaifeng 475004,Henan,China)

机构地区：[1]河南大学环境与规划学院,开封475004

出　　处：《地理研究》2018年第10期2058-2074,共17页Geographical Research

摘　　要：系统总结了空间数据统计分析的发展历程,并分为五个时期:(1)早期孕育(计量革命之前),其重要思想是19世纪初德国的区位论;(2)计量革命(1950-1960年代),主要是经典统计学的应用和理论探索;(3)空间统计学(1970-1980年代),重点是空间点数据、面数据和空间连续性数据的分析;(4)成熟与扩散(1990-2000年代),空间数据统计分析发展成熟并快速向其他领域扩散;(5)时空大数据(2010年以后)。换句话说,计量革命开始后的空间数据统计分析大约每20年有重要的新技术或方法出现,到现在已经具有成熟、系统化的方法和显著的社会效益。而在当前的时空大数据时期,其发展需要计算机科学家、统计学家和地理学家等不同学科领域人员的共同努力。Along with the historical background, characters and works of a particular period,this paper systematically summarizes the theory, method and technology of statistical analysis of spatial data（SASD）, and divides the SASD into five periods：（1） The early gestation（before the quantitative revolution）： Including German location theory in the early 19 th century, and the early studies in ecology, geology, etc.（2） Quantitative revolution（1950 s-1960 s）： Including mainly the direct application of classical statistics and mathematics, theoretical exploration, the understanding of spatial autocorrelation, and the birth of geostatistics.（3） Spatial statistics（1970 s-1980 s）： Including systematic research on spatial autocorrelation, and the analysis of spatial point data, lattice data, and spatial continuous data.（4） Maturation and diffusion（1990 s-2000 s）： With the help of computer, geographical information system（GIS） and spatial data collection technology, an in-depth study was conducted on large spatial databases and the spatial heterogeneity. It includes spatial data mining（SDM）, e.g., GeoMiner, and local spatial statistics such as local indicators of spatial autocorrelation（LISA）, geographical weighted regression（GWR）, spatial scan statistics, and GeoDetector. On the other hand, with the maturity and systematization of SASD, many works of summary and application in many fields have emerged naturally.（5） Spatio-temporal big data（2010 s and beyond）： This is the most important trend of SASD at present. In other words, since the quantitative revolution, SASD has produced important new methods or technologies every 20 years or so. In the current era of spatio-temporal big data, several research directions are worthy of attention, i.e., spatiotemporal point pattern and process, data streams analysis, network analysis, outlier detection,and uncertainty. In summary, after more than 60 years of development since quantitative revolution, SASD has become an

关 键 词：空间数据统计分析 空间自相关 空间统计学 空间数据分析 时空大数据 

分 类 号：F091[经济管理—政治经济学]