基于BP网络与空间统计分析的山东人口空间分布模式预测研究

运用BP网络对山东省17地市未来人口总量进行预测,在预测数据的基础上运用空间自相关方法对未来人口的空间分布模式进行分析。研究表明:2005-2010年山东省人口密度的空间分布模式总体呈现"西南-东北"模式,存在着空间集聚现象;2006-2010年17地市局部空间关联类型基本没有发生变化,西部和南部8个地区存在着明显的"高-高"集聚;北部5个地区存在着明显的"低-低"集聚;淄博和青岛存在着两个"高-低"关联的孤立点。

Prediction of population spatial distribution in Shandong based on BP ANN and spatial statistical analysis

Applying the BP artificial neural network, the population of 17 prefectures of Shandong province was predicted. Future population spatial distribution was analyzed using spatial autocorrelation. This study shows the general population density spatial distribution forms a southwest-northeast model. From the highest density in the southwest areas, the density decreases to the lowest density in the northeast areas, and the similar population density areas centralize in the vicinage areas. The partial spatial relation has not changed in 17 prefectures of Shandong province from 2006 to 2010. There are 8 "high-high" prefectures centralization in the west and south areas, there are 5 "low-low" prefectures centralization in the north areas, and there are 2 "high-low" relation isolated points in Zibo and Qingdao.