三值全局空间自相关指标及其应用研究

General Index of Spatial Autocorrelation with Trinary Variable and Its Application

Join-Count统计量是衡量空间自相关的全局指标。针对其用于处理二值名义变量的特性,将Join-Count统计量处理的属性值扩展到了三值变量,推导出了三值变量的邻接计数、邻接计数期望及方差的计算式,采用模拟数据进行了验算,证明所推导的公式是正确可行的,对计算结果的解释是合理的,这对三值Join-Count统计量应用于图像空间统计中的自相关分析有理论意义和实用参考价值。

Join-Count Statistic is the simplest index form of spatial autocorrelation, and it has ability to handle polygonal binary nominal data. The ability of processing binary nominal data to trinary nominal data is extended, the formulae of counting the joints is deduced, and the expectation and variance of the trinary join-count are calculated. At last, simulating data in image are used to testify the validity of the formulae.