摘要
模糊聚类分析方法在现实生活中被广泛应用,特别是模糊C均值算法(FCM)适用于很多场合。分析了传统FCM算法的不足并对此进行改进,提出了一种新的改进的算法,综合考虑每个样本的邻近相关度和评估函数,形成新的算法流程。通过对不同实验数据的应用,该算法具有较高的分类识别率和排除噪声能力。
The fuzzy clustering analysis method is widely used in real life,especially the fuzzy C means algorithm(FCM)is applicable to many occasions.This paper analyzes the shortcomings of the traditional FCM algorithm and improves it.A new improved algorithm is proposed,which takes into consideration the proximity correlation and evaluation function of each sample,and a new algorithm flow has been formed.Through the application of different experimental data,it got the conclusion that this algorithm has higher classification recognition rate and noise elimination ability.
出处
《长春工程学院学报（自然科学版）》
2018年第2期96-98,共3页
Journal of Changchun Institute of Technology：Natural Sciences Edition
基金
安徽省高校自然科学重点项目(KJ2015A402)
关键词
模糊聚类
模糊C均值
隶属度
评估函数
fuzzy clustering
fuzzy C means
membership degree
evaluation function
