基于改进样本构造方法的城市人口密度分布研究  

Spatial Distribution of Urban Population Density Based on Improved Method of Sample Construction

作　　者：康停军[1] 张新长[1] 赵元[2] 孙颖[1] 王海鹰 

机构地区：[1]中山大学地理科学与规划学院,广东广州510275 [2]华南农业大学信息学院,广东广州510642

出　　处：《地理与地理信息科学》2012年第2期50-54,共5页Geography and Geo-Information Science

基　　金：国家自然科学基金项目(40971216;41071246)

摘　　要：针对人口密度理论模型拟合时传统样本数据构造方法构造样本存在的问题,发展了一种多行政等级单元样本构造方法。利用广州市2000年人口统计数据及行政界线数据,构造单一行政等级和多行政等级单元样本;利用人口密度单中心模型和多中心模型分别进行了拟合实验,对拟合参数及拟合优度进行了分析。多行政等级单元样本在人口密度单中心模型拟合时既具有较好的拟合优度,又可以有效减轻偏大估计;在进行人口密度多中心拟合时可以发现近郊地区的副中心,能更细致、真实地描述人口分布多中心结构。The spatial distribution of population density is the starting point of studying social economy;it has arisen as an important issue in the fields of geographical and relative researches.Aiming at the problems of traditional sample construction methods for theoretical population mathematical models,this paper presented an improved method of sample named Multi-Administration Cells（MAC） method.Single Administration Cells（SAC） samples and MAC samples are constructed by using the 2000 population census and the administrative boundaries of Guangzhou.Then regression of monocentric and polycentric mathematical models for SAC and MAC samples is used to study the advantage of MAC samples.The authors have analyzed the regression parameters and goodness of fit for the entire regression models.Compared with the traditional sample construction method,the regression for monocentric models based on MAC samples not only have better goodness of fit but also reduce the upward bias of SAC samples,the regression for polycentric models based on MAC samples can find the sub-civic center that SAC samples has not found and can depict the polycentric structure of Guangzhou in 2000 in accord with actual conditions.

关 键 词：人口密度 单中心 多中心 偏大估计 广州 

分 类 号：C922[社会学—人口学]