地物分布的集散情形是空間現象重要的研究議題,部份度量集散的指標是以空間自相關為基礎,如Moran's I和Getis's G,部份指標則以其它方式加以權衡,如最近鄰法以最近鄰距離為主要依據,但這些指標都有一些侷限,因此,本研究嘗試建立一個新的指標:空間離散指標(Spatial Dispersion Index, SDI),該指標承襲統計學的離散觀念與公式,以統計單元大小、地物相對大小、鄰近程度高低為判定離散程度的三個原則,並以此來推導SDI的公式,該指標能應用在點資料與面資料的計算,但其最大值和數值分佈常因例而異,在判定SDI的集散意義時,部份方式計算較為複雜。
One of the controversial issues ill studying a spatial phenomenon is whether its distribution pattern is aggregated or dispersed. Some of the indexes are based on spatial autocorrelation, such as Moran's I and Getis's G Some are not, such as Nearest Neighbor Index is calculated with nearest neighbor distance. However, all of them have some limitations. The present research tries to develop a new index: Spatial Dispersion Index (SDI), following the concept and formula of dispersion from statistics. Three principles of SDI are adhered to: the area of statistical unit, comparative area of surface features and distance to neighbor. The SDI formula is deduced from those three principles. This index can calculate point and polygon features. But the SDI maximum value and value distribution will change by case. Estimating whether the SDI represents an aggregative or dispersive phenomenon, the calculations are more complex than other indexes.