引用本文:周洁,姜志彬,张远鹏,等.基于密度的模糊代表点聚类算法[J].控制与决策,2020,35(5):1123-1133
【打印本页】   【HTML】   【下载PDF全文】   查看/发表评论  【EndNote】   【RefMan】   【BibTex】 附件
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览次   下载 本文二维码信息
码上扫一扫!
分享到: 微信 更多
基于密度的模糊代表点聚类算法
周洁,姜志彬,张远鹏,王士同
( 江南大学数字媒体学院,江苏无锡214122;江苏省媒体设计与软件技术重点实验室,江苏无锡214122)
摘要:
结合密度聚类和模糊聚类的特点,提出一种基于密度的模糊代表点聚类算法.首先利用密度对数据点成为候选聚类中心点的可能性进行处理,密度越高的点成为聚类中心点的可能性越大;然后利用模糊方法对聚类中心点进行确定;最后通过合并聚类中心点确定最终的聚类中心.所提出算法具有很好的自适应性,能够处理不同形状的聚类问题,无需提前规定聚类个数,能够自动确定真实存在的聚类中心点,可解释性好.通过结合不同聚类方法的优点,最终实现对数据的有效划分.此外,所提出的算法对于聚类数和初始化、处理不同形状的聚类问题以及应对异常值等方面具有较好的鲁棒性.通过在人工数据集和UCI真实数据集上进行实验,表明所提出算法具有较好的聚类性能和广泛的适用性.
关键词:  聚类  密度聚类  模糊聚类  代表点聚类  聚类中心  鲁棒性
DOI:10.13195/j.kzyjc.2018.1179
分类号:TP181
基金项目:国家自然科学基金项目(61170122,61272210,81701793);江苏省自然科学基金项目(BK20130155);南通市科技计划项目(MS12017016-2).
A density-based fuzzy exemplar clustering algorithm
ZHOU Jie,JIANG Zhi-bin,ZHANG Yuan-peng,WANG Shi-tong
(School of Digital Media,Jiangnan University,Wuxi 214122,China;Jiangsu Key Laboratory of Digital Design and Software Technology,Wuxi 214122,China)
Abstract:
According to the characteristics of density-based clustering and fuzzy clustering, a density-based fuzzy exemplar clustering algorithm is proposed. Firstly, the possibility of data points becoming candidate clustering centers is processed by the density. The higher the density of the data point is, the greater the likelihood for the data point to become a clustering center is. The clustering centers are then selected using the fuzzy method. The final clustering centers are determined by merging the clustering centers. The proposed algorithm has great adaptability, which can deal with clustering problems of different shapes, it can not only automatically determine cluster centers, but also get better results with higher accuracy. It can automatically determine the real clustering centers with good interpretability and there is no need to preset the number of clusters in advance. By combining the advantages of different clustering methods, the effective division of data can be realized. In addition, it has better robustness to number of clusters and initialization, processing clustering problems of different shapes, and dealing with outliers. Experiments on synthetic datasets and UCI datasets show that the proposed algorithm has better clustering performance and wide applicability.
Key words:  clustering  density clustering  fuzzy clustering  exemplar clustering  clustering centers  robustness

用微信扫一扫

用微信扫一扫