一种非噪声敏感性的模糊C均值聚类算法  

An Insensitive to Noisy Fuzzy C-means Clustering Algorithm

作　　者：陈加顺[1,2] 皮德常[1] 

机构地区：[1]南京航空航天大学计算机科学技术学院,南京210016 [2]淮工淮海工学院计算机工程学院,江苏连云港222003

出　　处：《小型微型计算机系统》2014年第6期1427-1431,共5页Journal of Chinese Computer Systems

基　　金：国家"八六三"高技术研究发展计划项目(2007AA01Z404)资助;江苏省普通高校研究生科研创新计划项目(CXLX11_0206)资助

摘　　要：FCM算法提出用模糊隶属度表示样本数据的隶属于某个类的程度,能够克服HCM算法划分的不合理性.研究分析发现FCM算法对噪声数据具有敏感性,很难有效的识别噪声数据;FCM由于限制条件使得聚类结果与实际的分类不一致.针对此不足之处,文章提出一种非噪声敏感性FCM算法(INFCM),取消了限制条件,用典型值代替了隶属度值,构建了目标方程,为了克服聚类过程中一致性,在目标方程中增加了惩罚因子,分析了惩罚因子的组成,最后提出了聚类算法步骤.实验表明新的聚类算法能够有效克服对噪声的敏感性,提高了聚类的可理解性.Fuzzy C-means overcomes defects of hard C-means about unreasonable partition of samples and uses fuzzy membership to stand for some sample belonging to one class. Experiments show that Fuzzy C-mean is sensitive to noisy and difficult to recognize noisy data effetely. Another weak point is that clustering results are inconsistent with classification of real data. Aim at these deficiencies, we present an improved Fuzzy C means named INFCM （ Insensitive to Noisy Fuzzy C-means Clustering Algorithm）, and abolish constraints, and replace membership with typical value to construct objective function. In order to overcome coincident clustering results, we add penalty factor to objective function and analyze penalty factor. Finally, we present steps of algorithm. Experiments on different datasets show that improved algorithm has good performance and is insensitive to noisy, which enhance our intelligibility for clustering results.

关 键 词：模糊C均值 隶属度 惩罚因子 非噪声敏感性 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]