《计算机应用研究》|Application Research of Computers

基于LPCA的谱聚类算法

Spectral clustering algorithm based on LPCA

免费全文下载 (已被下载 次)  
获取PDF全文
作者 童涛,文国秋,谭马龙,吴林,杜婷婷
机构 广西师范大学 广西多源信息挖掘与安全重点实验室,广西 桂林 541004
统计 摘要被查看 次,已被下载
文章编号 1001-3695(2019)11-009-3245-05
DOI 10.19734/j.issn.1001-3695.2018.04.0283
摘要 针对传统谱聚类在构建关系矩阵时只考虑样本的全局特征而忽略样本的局部特征、在聚类划分时通常需要指定聚类个数、无法对交叉点进行正确划分等问题,提出了一种改进的基于局部主成分分析和连通图分解的谱聚类算法。首先自动学习挑选数据集的中心点,然后使用局部主成分分析得到数据集的关系矩阵,最后用连通图分解算法完成对关系矩阵的划分。实验结果表明该改进算法性能优于现有经典算法。
关键词 局部主成分分析; 谱聚类; 连通图分解; 交叉点
基金项目 国家重点研发计划资助项目(2016YFB1000905)
国家自然科学基金资助项目(61170131,61263035,61573270,90718020)
国家“973”计划资助项目(2013CB329404)
中国博士后科学基金资助项目(2015M570837)
广西自然科学基金资助项目(2015GXNSFCB139011,2015GXNSFAA139306)
广西研究生教育创新计划项目(YCSW2019072)
本文URL http://www.arocmag.com/article/01-2019-11-009.html
英文标题 Spectral clustering algorithm based on LPCA
作者英文名 Tong Tao, Wen Guoqiu, Tan Malong, Wu Lin, Du Tingting
机构英文名 Guangxi Key Laboratory of Multi-source Information Mining & Security,Guangxi Normal University,Guilin Guangxi 541004,China
英文摘要 As the traditional spectral clustering algorithms not only considered the global structures of the samples while ignoring their local structures for the construction of the correlation matrix, and conducted clustering with a predefined cluster number, but also could not divide the intersections correctly. This paper proposed a new method based on the local principal component analysis and the decomposition method of the connected graph. Specifically, the proposed method automatically learnt the centroids of the selected subset of the samples, obtained the correlation matrix of the samples based on the local principal component analysis, and used the decomposition method of the connected graph to partition the resulting correlation matrix. Experimental results show that the proposed algorithm performs better than the existing algorithms.
英文关键词 local principal component analysis(LPCA); spectral clustering; connected graph decomposition; intersection
参考文献 查看稿件参考文献
 
收稿日期 2018/4/17
修回日期 2018/5/28
页码 3245-3249
中图分类号 TP182
文献标志码 A