论文部分内容阅读
进行空间自相关分析时,由于权重矩阵构建的主观性和选择单一时间截面数据随机性影响,存在可能的选择偏倚性问题。扫描统计量方法则有效地解决上述问题。研究选择甘肃省中南部51个区、县,采用Geoda0.9.5和Clusterseer0.2.3软件,对2001～2010年卫生部监测点上报的30 463例围产儿死亡病例进行空间自相关和扫描统计量分析,从不同尺度上对分析结果进行验证,证明时空扫描统计量方法有显著的比较优势。
In the context of spatial autocorrelation analysis, there is a potential choice bias due to the subjectivity of the weight matrix construction and the randomness of selecting the single-time cross-sectional data. Scanning statistics methods effectively solve the above problems. This study selected 51 districts and counties in central and southern Gansu province. Using Geoda0.9.5 and Clusterseer0.2.3 software, the spatial autocorrelation and scanning statistic analysis of 30 463 cases of perinatal deaths reported by Ministry of Health monitoring points from 2001 to 2010 were conducted. The validation of the analytical results from different scales proves that spatio-temporal scanning statistics methods have significant comparative advantages.