障碍约束下的空间聚类是空间数据挖掘研究领域中一个重要的研究课题。论文研究了障碍约束的处理方法,引入粒子逃逸原则以避免聚类中心点陷入障碍物中,提出一种基于量子粒子群的绕过障碍物的空间聚类算法(QCOD),通过实验对比分析,该算法不仅有效地克服了划分聚类算法极易陷入局部极小值和对初始值敏感的问题,而且聚类结果比带障碍的k-中心点算法更符合实际情况。%Spatial clustering with obstacle contraints is one of the important areas of research projects in spatial data mining. This paper investigates the method of handling obstacle constraints, introduces the Escaping Principle to avoid the updated cluster center particles sinking into the area of the obstacles, and proposes a novel spatial clustering algorithm QCOD based on QPSO with obstacles constraints. The proposed method through the experimental contrast analysis effectively overcome the problems of easily falling into local extremum and sensitive to the initial parameters, and also it is better than k - mediods algorithm with obstacle constraints.
展开▼