吉林大学学报(信息科学版) ›› 2024, Vol. 42 ›› Issue (3): 476-485.

• • 上一篇    下一篇

高维矩阵奇异值分解的快速计算方法对比分析及应用

陈怡君1 , 韩 迪2 , 刘 骞3 , 徐海强3 , 曾海嫚2   

  1. 1. 西安航空学院 图书馆, 西安 710077; 2. 广东金融学院 信用管理学院, 广州 510521; 3. 西安交通大学 数学与统计学院, 西安 710049
  • 收稿日期:2023-04-26 出版日期:2024-06-18 发布日期:2024-06-17
  • 通讯作者: 韩迪(1982— ), 男, 武汉人, 广东金融学院副教授, 硕士生 导师, 主要从事推荐系统和机器学习研究, (Tel)86-13926108207(E-mail)dihan@ gduf. edu. cn E-mail:dihan@ gduf. edu. cn
  • 作者简介:陈怡君(1984— ), 女, 宁夏固原人, 西安航空学院副研究馆员, 主要从事图像识别和信息检索研究, ( Tel) 86- 13895690910(E-mail)201907034@ xaau. edu. cn
  • 基金资助:
     教育部人文社会科学研究规划基金资助项目(23YJAZH046)

Comparative Analysis and Application of Fast Calculation Methods for Singular Value Decomposition of High Dimensional Matrix

CHEN Yijun 1 , HAN Di 2 , LIU Qian 3 , XU Haiqiang 3 , ZENG Haiman 2   

  1. 1. Library, Xi’an Aeronautical Institute, Xian 710077, China; 2. School of Credit Management, Guangdong University of Finance, Guangzhou 510521, China; 3. School of Mathematics and Statistics, Xian Jiaotong University, Xian 710049, China
  • Received:2023-04-26 Online:2024-06-18 Published:2024-06-17

摘要: 为在大数据环境下处理高维矩阵和应用奇异值分解提供更高效的解决方案, 从而加速数据分析和处理速度, 通过研究随机投影以及 Krylov 子空间投影理论下关于高维矩阵求解特征值特征向量(奇异值奇异向量)问题, 分别总结了 6 种高效计算方法并对其相关应用研究进行对比分析。 结果表明, 在谱聚类的应用上, 通过降低核心步骤 SVD(Singular Value Decomposition)的复杂度, 使优化后的算法与原始谱聚类算法的精度相近, 但大大缩短了运行时间, 1 200 维的数据下计算速度相较原算法快了10 倍以上。 同时, 该方法应用于图像压缩领域, 能有效地提高原有算法的运行效率, 在精度不变的情况下, 运行效率得到了 1 ~ 5 倍的提升。

关键词:  高维矩阵, 快速奇异值分解, 谱聚类, 图像压缩

Abstract: To provide more efficient solutions for handling high-dimensional matrices and applying SVD(Singular Value Decomposition) in the context of big data, with the aim of accelerating data analysis and processing, how to quickly calculate the eigenvalues and eigenvectors ( singular value singular vectors) of high-dimensional matrices is studied. By studying random projection and Krylov subspace projection theory, six efficient calculation methods are summarized, making comparative analysis and related application research. Then, the six algorithms are applied, and the algorithms in related fields are improved. In the application of spectral clustering, the algorithm reduces the complexity of the core step SVD( Singular Value Decomposition), so that the optimized algorithm has similar accuracy to the original spectral clustering algorithm, but significantly shortens the running time. The calculation speed is more than 10 times faster than the original algorithm. When this work is applied in the field of image compression, it effectively improves the operation efficiency of the original algorithm. Under the condition of constant accuracy, the operation efficiency is improved by 1 ~ 5 times.

Key words: high-dimensional matrices, fast singular value decomposition ( SVD ), spectral clustering, image compression

中图分类号: 

  • TP391