使用python numpy進行SVD分解

>>> import numpy as np

>>> a=np.matrix([[3,0,3],[5,4,0],[1,2,4],[2,2,0]])

>>> a

matrix([[3, 0, 3],

        [5, 4, 0],

        [1, 2, 4],

        [2, 2, 0]])

>>> from numpy import linalg as la

>>> U,sigma,VT=la.svd(a)

>>> U

matrix([[-0.415103  , -0.4754462 , -0.76791353,  0.10931247],

        [-0.74371645,  0.52779221,  0.01689327, -0.40992176],

        [-0.41095874, -0.66262863,  0.62073729, -0.08198435],

        [-0.3250973 ,  0.23729643,  0.15718985,  0.90182787]])

>>> sigma

array([ 8.02645174,  4.38855108,  2.07766502])

>>> VT

matrix([[-0.75064858, -0.55404032, -0.35995283],

        [ 0.2334681 ,  0.28722565, -0.92897474],

        [-0.61807715,  0.78137107,  0.08625479]])

>>>