基于层次Voronoi图的点群相似度算法

Algorithm of point cluster similarity based on hierarchical Voronoi diagrams

通过对空间点群的自适应聚类方法构建层次Voronoi图,以此层次Voronoi图为切入点,计算点群的拓扑、密度和范围的相似度,结合有关标准差的数理统计方法,计算角度、距离的相似度。在各维度的相似度基础上,使用其几何平均值作为点群整体相似度的度量标准,优化点群相似度的计算方法,并通过实验证明算法的可行性。

The hierarchical Voronoi diagrams were built through an adaptive clustering method of spatial point clusters. Based on the hierarchical Voronoi diagrams, the topology, density and scope similarities were calculated. The radian and distance similarity were calculated in combination of the standard deviation in mathematical statistics. On the base of every dimensional similarity, the principle of point cluster similarity was decided by the geometrical mean of these parameters. This optimizes the method of the point cluster similarity and the experiment proves its feasibility.