交通小区的两维图论聚类

为使得交通小区合并生成交通中区的过程更加合理,同时考虑交通小区之间的相似性和位置关系,将两维图论聚类法应用于交通小区的合并.给定交通小区相邻满足的条件,并用邻接矩阵表示交通小区之间的位置关系,构造无向加权图并求解最小支撑树,根据最小支撑树选取阈值进行交通小区合并,最后用F检验法确定理论上的最优合并结果作为小区合并结果选取的参考.实例分析结果表明:聚类数随阈值的增大而减少,而且合并过程中只有相似且相邻的交通小区被合并,并采用F检验法确定了唯一的最优合并参考方案,划分结果合理可行.

Two-dimensional graphic theory on clustering method of small traffic zones

In order to make the process of generating small traffic zones into middle traffic zones more reasonable and taking the similarity and the positional relation of small traffic zones into consideration, two-dimensional graphic theory-clustering method on the mergence of small traffic zone is applied. Given conditions that satisfied the adjacent small traffic zones, the positional relation of small traffic zones by the adjacent matrix is re-formulated. After the undirected weighted graph construction, the minimum spanning tree (MST) was solved. Choose the threshold according to the MST for the small traffic zones mergence. Finally the F-test is used to determine the optimal mergence in theory which is the reference of the selection of the small traffic zone mergence. The result of example analysis shows that the clustering number reduces along with the increase of the threshold value, and only the similar and adjacent small traffic zones can be merged during the mergence. In addition, the F-test is adopted to determine the exclusive optimal scheme of mergence for reference, which proved that the division result was feasible.