江苏省科技期刊可视化阅读平台

导航条

《无线互联科技》杂志社 ›› 2022, Vol. 19 ›› Issue (9): 147-149.

• 实验研究 • 上一篇    下一篇

基于流行学习的降维算法研究

陈小军   

  1. 贵阳银行股份有限公司,贵州 贵阳 550081
  • 出版日期:2022-05-10 发布日期:2022-07-25
  • 作者简介:陈小军(1982— ),男,贵州黔东南人,中级经济师,高级信息系统项目管理师,硕士;研究方向:数据分析,数据挖掘,人工智能,机器学习,深度学习。

Research on dimension reduction algorithm based on manifold learning

Chen Xiaojun   

  1. Guiyang Bank Co., Ltd., Guiyang 550081, China
  • Online:2022-05-10 Published:2022-07-25

PDF

791

摘要： 流行学习一直是深度学习、机器学习、人工智能的研究热点,其是非线性降维算法中非常重要算法思想。流行学习利用局部与欧式空间是同胚空间的概念,将高维数据空间嵌入低维流行结构,用低维局部欧氏特征表示高维数据结构。文章介绍了流行学习的基本概念、经典算法思想,对经典算法思想进行分析比较,并提出有待进一步分析和研究的问题。

关键词: 降维算法, 非线性降维, 流行学习, 机器学习

Abstract: Manifold learning has always been a research hotspot in deep learning, machine learning and artificial intelligence. Manifold learning is a very important algorithm idea in nonlinear dimensionality reduction algorithms, which uses the characteristics of homeomorphic space between local and euclidean space to embed high-dimension data space in low dimensional manifold structures in low dimensions. This article introduces the basic ideas of manifold learning, classic algorithm ideas and the shortcomings of classic algorithms, which Analysis and comparison in the article,puts forward questions that need further analysis and research.

Key words: imensionality reduction algorithm, non-linear dimensionality, manifold learning, machine learning