蚁群聚类组合方法参数M的研究.doc

蚁群聚类组合方法参数M的研究
　　摘要:蚁群算法中参数在不同取值情况下,常常会对算法的性能和求解效率产生重大影响。该文在基于蚁群聚类组合方法的研究基础上,重点研究了蚁群聚类组合方法KMAOC算法中蚁群算法参数蚂蚁数m对KMAOC算法性能的影响,对KMAOC算法中的参数蚂蚁数m分别取值进行实验,通过几组实验验证提供了KMAOC算法中参数蚂蚁数m配置的较好建议。
　　关键词:聚类;蚁群算法;信息素;聚类组合
　　中图分类号:TP311文献标识码:A文章编号:1009-3044(2009)36-10469-03
　　Research Combination Method of Parameter M Based on Ant Colony Clustering
　　HAN Qiang, XING Jie-qing
　　(Department of Modern Education Technology, Qiongtai Teachers College, Haikou 571100, China)
　　Abstract: The ant-based clustering parameter values in different circumstances, often will solve the performance and efficiency of the algorithm have a significant impact. In this paper, based on ant colony clustering combination method based on the study, focusing on the ant colony clustering algorithm combination m
ethod KMAOC ant colony algorithm parameters are the number of m pairs of KMAOC algorithm performance influence on the parameters of the algorithm KMAOC the number of ants m, respectively experimental values by several groups of experimental verification provides the better proposal that a KMAOC ant algorithm parameters to configure the number of m.
　　Key words: clustering; ant colony algorithm; pheromone; clustering combination
　　聚类在科学数据探测、图像处理、模式识别、医疗诊断、计算生物学、文档检索、Web分析等领域起着非常重要的作用,它已经成为当前数据挖掘研究领域中一个非常活跃的研究课题[1]。经典聚类方法包括分层算法,划分方法如K均值算法、模糊C均值算法,图论聚类法,神经网络法,以及基于统计的方法等[2]。近来随着数据挖掘研究的深入,涌现了大量新的聚类算