摘要
为解决聚类数未知条件下面状地理实体的聚类问题,文中提出了一种基于聚类有效性函数的聚类方法。给出了适合面状地理实体k-中心点聚类算法的聚类有效性函数;将该有效性函数改写为适应度函数,设计了基于遗传算法的面状地理实体聚类算法。该算法在计算聚类数的同时能得到划分聚类结果。实验结果从一定程度上反映了数据集的结构信息特征。
A cluster validity function-based method is proposed for solving the problem of clustering for area geographical entities when the number of cluster is unknown. A cluster validity function fitting to k-medoid clustering algorithm is derived and is taken as fitness function. A GAs-based clustering algorithm is put forth. The advantage of the algorithm is getting clusters of dataset as well as calculation the number of cluster. The results of test reflect the inner properties of dataset in some extent.
出处
《测绘科学技术学报》
北大核心
2006年第1期44-47,共4页
Journal of Geomatics Science and Technology
关键词
聚类有效性函数
遗传算法
聚类
cluster validity function
genetic algorithms
clustering
