摘要
在分析、验证主流降维算法性能的基础上,设计了基于聚类的分层降维框架,将聚类和降维结合,实现了类内和类间分别降维的处理机制。实验结果表明,随着数据集规模增长,分层降维框架的时间效率逐渐体现出优势。
By analyzing and verifying the performance of mainstream dimension reduction algorithms,a hierarchical dimension reduction framework based on clustering algorithms was proposed,which combined clustering and dimensionality reduction to realize the processing mechanism of dimensionality reduction within and between classes respectively.Experimental results of time efficiency show that this framework bears some advantages over the original dimension reduction scheme as data volume increases.
作者
陈新元
谢晟祎
CHEN Xin-yuan;XIE Sheng-yi(College of Computer and Control Engineering,Minjiang University,Fuzhou 350121,China;Department of Information Engineering,Fuzhou Melbourne Polytechnic,Fuzhou 350121,China;Experimental Training Center,Fujian Vocational College of Agriculture,Fuzhou 350181,China)
出处
《唐山师范学院学报》
2020年第3期78-82,共5页
Journal of Tangshan Normal University
基金
福建省中青年教师教育科研项目(JAT160316)。
关键词
高维数据
降维
聚类
分层
分类准确率
multiple dimensional data
dimension reduction
clustering
stratification
classification accuracy
