摘要
聚类分析的很多算法中都采用连接两点直线的欧氏距离来判断空间亲疏性,然而当有障碍物层存在时,连接两点的直线已不能完全表达它们之间的关系,特别是当有指定的交通路线时,两点之间的连通路径和距离必须遵从特定的路径。文中讨论在了障碍物或指定的交通路线存在的情况下点集的聚类分析,给出了解决有障碍物或指定交通路线时进行聚类分析的算法PathClust。
Many clustering algorithms use the Euclidean However, when the layers of obstacles presented, the tions completely. Especially, when the transportation trace the some fixed path. PathClust is put forward to distance of two points to measure the spatial proximity, line linking the two points could not represent the rela network exists, the route linking the two points must realize the clustering of points in the presence of obstacles or the transportation networks.
出处
《测绘科学技术学报》
北大核心
2006年第2期145-148,共4页
Journal of Geomatics Science and Technology
关键词
聚类分析
地址匹配
最佳路径
最佳距离
clustering analysis
address geoeoding
optimum route
the shortest distance