南方人口
南方人口
남방인구
SOUTH CHINA POPULATION
2014年
3期
1-9
,共9页
复杂地形区%空间依赖%人口分布悖论%空间计量模型%影响因子
複雜地形區%空間依賴%人口分佈悖論%空間計量模型%影響因子
복잡지형구%공간의뢰%인구분포패론%공간계량모형%영향인자
Region of complex terrain%Spatial dependence%Paradox of population distribution%Spatial econometric model%Influent factors%GIS
提出并评述了关于人口分布传统研究的三个衍生问题：空间尺度不同可能产生“人口分布悖论”现象、空间相互独立的假设不尽合理、特定空间区域人口分布的复杂性。以此为切入点，通过空间计量分析方法，探讨中国川西高原高山区人口分布特征。研究表明，全局Moran's I值为0．7404，指示川西复杂地形区人口分布有十分显著的集群特征。以川西48个县域单元为对象，建立人口密度与相关影响因子的OLS估计、 SLM估计和SEM估计，经空间依赖性检验，发现空间计量回归模型明显优于普通最小二乘回归模型，表明考虑了空间依赖性的估计方法能更好的刻画人口空间分布特征；经相关判别检验，确立SLM为最优模型，其纠正了OLS模型总体上高估自然环境因子而低估经济社会因子对人口密度影响的缺陷。并以SLM估计结果为基础分析了川西人口分布特征，特别解析了川西“人口分布悖论”的事实和原因。
提齣併評述瞭關于人口分佈傳統研究的三箇衍生問題：空間呎度不同可能產生“人口分佈悖論”現象、空間相互獨立的假設不儘閤理、特定空間區域人口分佈的複雜性。以此為切入點，通過空間計量分析方法，探討中國川西高原高山區人口分佈特徵。研究錶明，全跼Moran's I值為0．7404，指示川西複雜地形區人口分佈有十分顯著的集群特徵。以川西48箇縣域單元為對象，建立人口密度與相關影響因子的OLS估計、 SLM估計和SEM估計，經空間依賴性檢驗，髮現空間計量迴歸模型明顯優于普通最小二乘迴歸模型，錶明攷慮瞭空間依賴性的估計方法能更好的刻畫人口空間分佈特徵；經相關判彆檢驗，確立SLM為最優模型，其糾正瞭OLS模型總體上高估自然環境因子而低估經濟社會因子對人口密度影響的缺陷。併以SLM估計結果為基礎分析瞭川西人口分佈特徵，特彆解析瞭川西“人口分佈悖論”的事實和原因。
제출병평술료관우인구분포전통연구적삼개연생문제：공간척도불동가능산생“인구분포패론”현상、공간상호독립적가설불진합리、특정공간구역인구분포적복잡성。이차위절입점，통과공간계량분석방법，탐토중국천서고원고산구인구분포특정。연구표명，전국Moran's I치위0．7404，지시천서복잡지형구인구분포유십분현저적집군특정。이천서48개현역단원위대상，건립인구밀도여상관영향인자적OLS고계、 SLM고계화SEM고계，경공간의뢰성검험，발현공간계량회귀모형명현우우보통최소이승회귀모형，표명고필료공간의뢰성적고계방법능경호적각화인구공간분포특정；경상관판별검험，학립SLM위최우모형，기규정료OLS모형총체상고고자연배경인자이저고경제사회인자대인구밀도영향적결함。병이SLM고계결과위기출분석료천서인구분포특정，특별해석료천서“인구분포패론”적사실화원인。
This article first reviews the three derivative problems in the traditional research of population distribution, i.e. the paradox of population distribution with spatial scales,irrational hypothesis of spatial independence,complication of population distribution in aparticularregion and based on these discussion,the spatial econometric model is used to analyze the characteristics of population distributionin the plateau and alpine area of Western Sichuan in china. It finds that the global Moran's I is 0.7404 demonstrating a clearly significant cluster feature of population distribution in the Western Sichuan Region with complex terrain. The estimation of OLS, SLM and SEM of the population densities and its relevant influencing factors of the 48 counties and the test of spatial dependence shows that spatial econometric model is much better than OLS model, which means that the estimation methods with spatial dependence can better reflect the characteristic of population distribution. The relevant differentiation tests show that SLM is the best model, because it can correct the defect of OLS model to overestimate the influence of natural factor while to underestimating the influence of social economic factor shaping population density. Finally the characteristics of population distribution of Western Sichuan are analyzed by SLM model to explain the causes of the population distribution paradox there.
