一种空间权重矩阵的优化方法

发布时间：2018-05-02 16:47

本文选题：空间自相关 + 空间权重矩阵　；参考：《武汉大学学报(信息科学版)》2017年06期

【摘要】：在地理学空间自相关的分析中,权重矩阵对整个分析结果有着较大影响。常见的权重矩阵,例如车矩阵、皇后矩阵、距离权重矩阵和k-邻近矩阵,都有各自的优势和缺点。提出了一种基于长度面积比例的空间权重矩阵(ratio of length and area,RLA),并以近年来危害最大的几种传染病之一——病毒性肝炎在中国大陆各省份的发病率为例进行了实验分析。实验结果表明,RLA矩阵能够很好地实现空间权重矩阵的基本功能,是常见的车矩阵的一种更为广义的定义,并且可以更加自由地实现空间自相关的分析。利用本空间权重矩阵能够更好地模拟不同地理单元之间的邻接关系,为流行病的预防提供支持。
[Abstract]:In the analysis of spatial autocorrelation in geography, the weight matrix has a great influence on the whole analysis result. The common weight matrices, such as vehicle matrix, queen matrix, distance weight matrix and k-adjacent matrix, all have their own advantages and disadvantages. A spatial weight matrix ratio (of length and) based on length area ratio is proposed. The incidence of viral hepatitis, one of the most harmful infectious diseases in recent years, is analyzed experimentally in various provinces of mainland China. The experimental results show that the RLA matrix can realize the basic function of the spatial weight matrix, is a more generalized definition of the common vehicle matrix, and can realize the spatial autocorrelation analysis more freely. The local spatial weight matrix can better simulate the adjacency between different geographical elements and provide support for epidemic prevention.
【作者单位】：北京大学地球与空间科学学院;北京大学工学院;
【基金】：高分辨率对地观测系统国家重大专项(03-Y30B06-9001-13/15,30-Y30B13-9003-14/16) 测绘地理信息公益性行业科研专项(201512020)~~
【分类号】：R181

【相似文献】