优化的核模糊C均值聚类算法  

Optimized Kernel Fuzzy C-Means Clustering Algorithm

作　　者：刘奕麟 安建成[1] 

机构地区：[1]太原理工大学计算机科学与技术学院,山西太原030024

出　　处：《微电子学与计算机》2018年第2期79-83,共5页Microelectronics & Computer

基　　金：山西省国际科技合作项目(2014081018-2)

摘　　要：提出一种优化的核模糊C均值聚类算法(WBAKFCM).该算法首先通过改进蝙蝠算法(Weight bat Algorithm,WBA)确定最优聚类中心集合,然后用核模糊C均值聚类算法指导聚类划分.一方面,改进的蝙蝠算法在传统的蝙蝠算法中引入佳点集理论和速度权重,分别用于调节种群的初始化和个体位置的自适应更新.另一方面,在核模糊C均值聚类算法(Kernel Fuzzy C-Means,KFCM)中,选用了高斯核函数,从而将数据映射到高维特征空间进行聚类划分.实验结果表明,优化的核模糊C均值聚类算法在聚类准确率与时间效率上明显优于传统算法.This paper an optimized kernel fuzzy C-means clustering algorithm （WBAKFCM） is proposed. Firstly,the optimal clustering center is found by the improved bat algorithm （WBA）, then the Kernel Fuzzy C-Meansclustering algorithm （KFCM） is used to guide the clustering. On the one hand, the improved bat algorithm addingtwo strategies to the traditional bat algorithm, the good point set theory and velocity weight are used to adjustpopulation initialization and adaptive updates of the individual position respectively. On the other hand, in theKernel Fuzzy C-Means clustering algorithm, the Gaussian kernel function is selected to map the data to high-dimensional feature space for clustering. The experimental results show that the optimized kernel fuzzy C-meansclustering algorithm is superior to the traditional algorithm in clustering accuracy and time efficiency.

关 键 词：模糊C均值聚类 核函数 蝙蝠算法 佳点集 速度权重 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]