摘要
将克隆选择原理同典型的划分聚类方法结合起来,提出一种克隆选择聚类算法.该算法具有完成任意形状数据集聚类的能力,可以自动确定簇的数目并得到簇的描述信息,计算量小,参数设置容易,适用于具有实值连续属性的数据集.基于模拟数据集和基准数据集分别进行实验,结果表明该算法是有效的.
Based on clone selection theory and typical partition clustering approach, a new clustering algorithm is proposed. The algorithm is capable of completing clustering of datasets with arbitrary shape, and acquiring the number and description of clusters. Besides, the algorithm has smaller computational cost, and the parameter can be set easily. The algorithm is applicable to the datasets that have real-value and continuous attributes. The experiments with two simulation datasets and three benchmark datasets show the effectiveness of the algorithm.
出处
《控制与决策》
EI
CSCD
北大核心
2005年第11期1261-1264,共4页
Control and Decision
关键词
克隆选择
聚类算法
簇分析
数据集
Clone selection
Clustering algorithm
Cluster analysis
Datasets
