摘要
传统谱聚类算法经常在处理一些结构复杂的数据集时效果不太理想,并且其相似度矩阵构造时参数的选取往往需要依靠多次实验及个人经验。在这种情况下,提出一种基于自然最近邻相似图的谱聚类(NSG-SC)算法。自然最近邻是一种新颖的最近邻概念,可以有效地避免K最近邻以及ε-最近邻方法需要人为设置参数的缺点。该算法构造相似度矩阵时依靠数据集自身的特性进行搜索,避免了参数选取不当以及离散点所带来的影响,更加真实地反映了数据集的结构关系。实验结果表明,提出的NSG-SC算法具有可行性和有效性。
The traditional spectral clustering algorithm cannot often get correct results on complex data sets,and the choice of parameters of affinity matrix construction depends on multiple tests and personal experience.Based on the situation,this paper proposed a spectral clustering algorithm based on natural nearest neighbor similarity graph(NSG-SC).Natural nearest neighbor was a novel concept in terms of nearest neighbor,and it could avoid the disadvantages of K-nearest neighbor andε-nearest neighbor.They usually needed set parameters artificially effectively.The algorithm constructed an affinity matrix depending on the characteristics of the data sets,and it avoided some adverse effects.It was that inappropriate choice of parameters and isolated points cause them.The algorithm could also reflect better characteristics of data.The results of experiment show that the proposed algorithm named NSG-SC has feasibility and effectiveness.
作者
刘友超
张曦煌
Liu Youchao;Zhang Xihuang(School of Internet of Things Engineering,Jiangnan University,Wuxi Jiangsu 214122,China)
出处
《计算机应用研究》
CSCD
北大核心
2020年第1期30-33,39,共5页
Application Research of Computers
基金
江苏省产学研合作项目(BY2015019-30).
关键词
谱聚类
自然最近邻
相似图
相似度矩阵
spectral clustering
natural nearest neighbor
similarity graph
affinity matrix