基于随机森林模型的珠江三角洲30 m格网人口空间化
发布时间：2019-04-11 14:05
【摘要】：人口空间化是实现人口统计数据与其他环境资源空间数据融合分析的有效途径。本文选取夜间灯光数据、道路网数据、水域分布数据、建成区数据、数字高程模型和地形坡度数据作为影响珠江三角洲人口分布的变量因子,利用随机森林模型对珠江三角洲2010年人口数据进行了30 m格网空间化,并将模拟结果与三个公开数据集作精度对比,最后基于随机森林模型的变量因子重要性分析珠江三角洲人口空间分布的影响因素。结果表明:本文模拟整体精度达到82.32%,均优于World Pop数据集以及中国公里网格人口数据集,接近GPW数据集,而且在人口密度中等区域模拟精度最高;通过对变量因子重要性进行度量,发现夜间灯光强度是珠江三角洲人口分布的最重要指示性指标,到水域的距离、到建成区的距离和路网密度对珠江三角洲人口分布均具有重要作用。利用随机森林模型结合多源信息能够实现高空间分辨率的人口空间化,可为精细化城市管理提供重要数据源,也可为相关政策决策制定提供支持。
[Abstract]:The spatialization of population is an effective way to realize the fusion analysis of population statistics and other environmental resources spatial data. In this paper, night lighting data, road network data, water distribution data, built area data, digital elevation model (Dem) and topographic slope data are selected as variable factors affecting population distribution in the Pearl River Delta (PRD). The population data of Pearl River Delta (PRD) in 2010 are spatialized by random forest model, and the simulation results are compared with the accuracy of three open data sets. Finally, the influence factors of population spatial distribution in Pearl River Delta are analyzed based on the variable factor importance of stochastic forest model. The results show that the overall accuracy of the simulation is 82.32%, which is better than the World Pop data set and the Chinese kilometer grid population data set, and is close to the GPW data set. Moreover, the simulation precision is the highest in the medium area of population density. By measuring the importance of variable factors, it is found that nighttime light intensity is the most important indicator of population distribution in the Pearl River Delta, and the distance to the waters. The distance to the area and the density of the road network play an important role in the population distribution of the Pearl River Delta (PRD). The random forest model combined with multi-source information can realize the population spatialization with high spatial resolution, which can provide important data sources for fine urban management and support the decision-making of related policies.
【作者单位】： 中山大学地理科学与规划学院广东省城市化与地理环境空间模拟重点实验室综合地理信息研究中心;
【基金】：国家自然科学基金重点项目(41531178) 广州市科技计划项目(201510010081) 国家自然科学基金项目(41001291)~~
【分类号】：C924.1
，
本文编号：2456469
[Abstract]:The spatialization of population is an effective way to realize the fusion analysis of population statistics and other environmental resources spatial data. In this paper, night lighting data, road network data, water distribution data, built area data, digital elevation model (Dem) and topographic slope data are selected as variable factors affecting population distribution in the Pearl River Delta (PRD). The population data of Pearl River Delta (PRD) in 2010 are spatialized by random forest model, and the simulation results are compared with the accuracy of three open data sets. Finally, the influence factors of population spatial distribution in Pearl River Delta are analyzed based on the variable factor importance of stochastic forest model. The results show that the overall accuracy of the simulation is 82.32%, which is better than the World Pop data set and the Chinese kilometer grid population data set, and is close to the GPW data set. Moreover, the simulation precision is the highest in the medium area of population density. By measuring the importance of variable factors, it is found that nighttime light intensity is the most important indicator of population distribution in the Pearl River Delta, and the distance to the waters. The distance to the area and the density of the road network play an important role in the population distribution of the Pearl River Delta (PRD). The random forest model combined with multi-source information can realize the population spatialization with high spatial resolution, which can provide important data sources for fine urban management and support the decision-making of related policies.
【作者单位】： 中山大学地理科学与规划学院广东省城市化与地理环境空间模拟重点实验室综合地理信息研究中心;
【基金】：国家自然科学基金重点项目(41531178) 广州市科技计划项目(201510010081) 国家自然科学基金项目(41001291)~~
【分类号】：C924.1
，
本文编号：2456469
本文链接：https://www.wllwen.com/shekelunwen/renkou/2456469.html
