34.24 SVD 分解

Fast truncated singular value decompositions 奇异值分解是特征值分解的推广

svd(A)
## $d
## [1] 3.6502815 0.8218544
## 
## $u
##            [,1]       [,2]
## [1,] -0.1601822 -0.9870875
## [2,] -0.9870875  0.1601822
## 
## $v
##            [,1]       [,2]
## [1,] -0.5847103 -0.8112422
## [2,] -0.8112422  0.5847103
svd(A)$d
## [1] 3.6502815 0.8218544

邱怡轩将奇异值分解用于图像压缩 https://cosx.org/2014/02/svd-and-image-compression 并制作了 Shiny App 交互式演示