中国省际人口迁移空间特征与影响因素分析

发布时间：2018-02-26 00:30

本文关键词： 人口迁移流网络自相关特征向量空间过滤模型　出处：《华南师范大学学报(自然科学版)》2017年03期 　论文类型：期刊论文

【摘要】：采用全国第六次人口普查的省际人口迁移及相关社会经济等数据,首先分析省际迁移人口的空间分布特征,然后利用全局Moran's I指数考察了省际人口迁移流中的网络自相关性,再构建基于网络自相关的特征向量空间过滤模型对省际人口迁移的动力机制进行分析,并与引力模型的回归结果进行对比验证,揭示网络自相关影响下省际人口迁移的动力机制.结果表明:(1)省际迁入及净迁入人口主要集中在我国三大经济圈,省际迁出人口主要分布于我国中南部;省际总迁移人口积聚于三大经济圈及中南部地区.(2)省际人口迁入、迁出流存在网络自相关,对人口迁移动力建模时应考虑网络自相关因素.文中加入了网络自相关因素后的特征向量空间过滤模型的拟合水平整体优于引力模型,较成功地揭示了人口迁移流中的网络自相关效应,减少了对其他变量的有偏估计.(3)非网络自相关变量中,人口总量、经济及距离因素对人口迁移活动的影响较大.
[Abstract]:Based on the data of interprovincial population migration and related social and economic data from 6th national population censuses, the spatial distribution characteristics of interprovincial migration population are analyzed, and then the network autocorrelation in interprovincial population migration flow is investigated by using the global Moran's I index. Then we construct the eigenvector space filtering model based on network autocorrelation to analyze the dynamic mechanism of interprovincial population migration and compare it with the regression result of gravity model. The results show that the inter-provincial migration and net migration are mainly concentrated in the three economic circles of China, and the inter-provincial emigration population is mainly distributed in the central and southern part of China. The total interprovincial migration population accumulates in the three economic circles and the central and southern regions.) the interprovincial population migrates in and out of the province, and there is a network autocorrelation between the emigration flows. The autocorrelation factor of network should be taken into account in the dynamic modeling of population migration. The fitting level of feature vector space filtering model with network autocorrelation factor is better than that of gravity model as a whole. The network autocorrelation effect in population migration flow is successfully revealed, and the biased estimation of other variables is reduced. Among the non-network autocorrelation variables, the total population, economy and distance factors have a great influence on population migration activities.
【作者单位】：华南师范大学地理科学学院;
【基金】：国家自然科学基金项目(41001078)
【分类号】：C922

【相似文献】