核参数优化选取的混合C均值核模糊聚类算法  

Mercer-kernel based mixed C-means fuzzy clustering algorithm with optimized kernel parameters

作　　者：江南[1] 王士同[1] 贺杨成[1] 

机构地区：[1]江南大学数字媒体学院,江苏无锡214122

出　　处：《计算机工程与设计》2011年第9期3148-3152,共5页Computer Engineering and Design

基　　金：国家863高技术研究发展计划基金项目(2007AA1Z158);国家自然科学基金项目(60704047);国家自然科学基金重大研究计划基金项目(9082002)

摘　　要：为了减小噪声点对聚类中心的影响,可能性聚类算法(PCM)把可能隶属关系引入到聚类的过程中,但是其往往趋向找到相同的集群。PFCM算法同时利用隶属度与可能性把数据点划分到不同的集群中,提高了算法的抗噪能力,但PFCM算法对发现大小不相等的集群并不十分理想。针对上述问题,提出了一种核参数优化选取的混合C均值核模糊聚类算法,该算法通过利用优化选取核参数的核函数把在原始空间中非线性可分的群体转化成高维空间中同质集群。实验结果表明,该算法能更好地发现融入噪音数据集的聚类中心,获得数据集质量更好的划分。To reduce the noise＇s effect on cluster centers,a possibilistic clustering algorithm use possibilities in process of clustering,but it often tends to find the identical.To overcome this shortcoming,PFCM is proposed which divides the data set into k different clusters through producing memberships and possibilities simultaneously,along with the cluster centers.But when two highly unequal sized clusters are given,PFCM fails to give the desired results.Therefore,a mercer-kernel based mixed c-means fuzzy clustering algorithm with optimized kernel parameters is proposed,which uses kernel function and makes it possible to cluster data that is linearly non-separable in the original space into homogeneous groups in the transformed high dimensional space with optimized kernel parameters.The experimental results show that the proposed algorithm can precisely find the ideal cluster centers and gives a better results.

关 键 词：核 模糊聚类 模式识别 可能性 混合 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]