| 摘 要: 机器学习是近几年研究的热点，维数约简算法是机器学习的必要手段，本文从维数约简算法的定义讲 起，介绍了几种典型的数据降维算法，其中包括线性降维和非线性降维，流形学习是非线性降维的代表算法。并且介绍 了每个算法的构造过程及其特点，在此基础上分析了所有维数约简算法的执行效率时间和空间复杂度，并且给出了每个 算法的特点和算法的核心思想，最后在此基础上给予总结，为后面研究者提供参考和借鉴。 |
| 关键词: 机器学习；维数约简；数据降维；线性降维；非线性降维 |
|
中图分类号: TP301
文献标识码: A
|
| 基金项目: 国家自然科学基金项目(60472121)，商洛学院自然科学研究项目(15SKY007). |
|
| Introduction of the Dimensionality Reduction Algorithm |
|
MA Famin,ZHANG Lin,WANG Jinbiao1,2
|
1.( 1.Institute of Mathematics and Computer Application, Shangluo University, Shangluo 726000, China;2.
2.College of Computer Science and Technology, Civil Aviation University of China, Tianjin 300000, China)
|
| Abstract: Machine learning,mainly realized through dimensionality reduction,has become a hot topic for research in recent years.This paper first presents the definition of the dimensionality reduction algorithm,and then introduces several typical data dimensionality reduction algorithms including linear dimensionality reduction and non-linear dimensionality reduction(manifold learning is the typical algorithm of non-linear dimensionality reduction).Besides,the paper elaborates on the construction process and characteristics of each algorithm,then analyzes the execution efficiency time and space complexity of all dimensionality reduction algorithms and provides the features and key point of each algorithm.Most importantly,the final conclusion offers references to future researchers. |
| Keywords: machine learning;dimensionality reduction;data dimensionality reduction;linear dimensionality reduction; non-linear dimensionality reduction;manifold learning |
