基于K-means算法的改进蚁群聚类算法及其应用  

Improved AntClust Algorithm Based on K-means Algorithm and Its Application

作　　者：莫赞[1] 罗世雄[1] 杨清平[1] 吴静[1] 

机构地区：[1]广东工业大学管理学院,广东广州510520

出　　处：《系统科学学报》2012年第3期91-95,共5页Chinese Journal of Systems Science

基　　金：广东省科技计划项目(2008B021300015);广东省哲学社会科学"十一五"规划项目(07SJY020)

摘　　要：基于化学识别的蚁群聚类算法无需给定聚类数目就能自动实现数据集的聚集,但大量采用随机策略使得蚂蚁达到平衡的运行时间长,效率不高。为此提出了一种用K-means算法做初次聚类、蚁群聚类算法再次聚类的新算法,结果表明该算法具有较高的正确率。最后将该算法应用于系统应用协同中,得到了较好的分析结果。AntClust algorithm based on chemical identification does not need a given number of clus- ters can automatically gather data sets, but the strategy makes a lot of random ants running for a long time to reach equilibrium, the efficiency is not high. This paper presents a new algorithm that with the initial K -means clustering, and antclust algorithm for the second clustering, the results show that the new algo- rithm has higher accuracy. The algorithm is applied to Collaboration system applications to get better re- suits at last.

关 键 词：K-MEANS算法 改进蚁群聚类算法 系统应用协同 

分 类 号：N949[自然科学总论—系统科学]