Python中的矩阵操作

Numpy

通过观察Python的自有数据类型，我们可以发信啊Python原生并不提供多为数组的操作，那么为了处理矩阵，就需要使用第三方提供的相关的包。

NumPy 是一个非常优秀的提供矩阵操作的包。NumPy的主要目标，就是提供多维数组，从而实现矩阵操作。

NumPy’s main object is the homogeneous multidimensional array. It is a table of elements (usually numbers), all of the same type, indexed by a tuple of positive integers. In NumPy dimensions are called axes.

基本操作

#######################################
# 创建矩阵
#######################################
from numpy import array as matrix, arange


# 创建矩阵
a = arange(15).reshape(3,5)
a


# Out[10]:
# array([[0., 0., 0., 0., 0.],
#        [0., 0., 0., 0., 0.],
#        [0., 0., 0., 0., 0.]])


b = matrix([2,2])
b


# Out[33]: array([2, 2])


c = matrix([[1,2,3,4,5,6],[7,8,9,10,11,12]], dtype=int)
c


# Out[40]:
# array([[ 1,  2,  3,  4,  5,  6],
#        [ 7,  8,  9, 10, 11, 12]])

#######################################
# 创建特殊矩阵
#######################################
from numpy import zeros, ones,empty


z = zeros((3,4))
z


# Out[43]:
# array([[0., 0., 0., 0.],
#        [0., 0., 0., 0.],
#        [0., 0., 0., 0.]])


o = ones((3,4))
o


# Out[46]:
# array([[1., 1., 1., 1.],
#        [1., 1., 1., 1.],
#        [1., 1., 1., 1.]])


e = empty((3,4))
e


# Out[47]:
# array([[0., 0., 0., 0.],
#        [0., 0., 0., 0.],
#        [0., 0., 0., 0.]])

#######################################
# 矩阵数学运算
#######################################
from numpy import array as matrix, arange


a = arange(9).reshape(3,3)
a


# Out[10]:
# array([[0, 1, 2],
#        [3, 4, 5],
#        [6, 7, 8]])


b = arange(3)
b


# Out[14]: array([0, 1, 2])


a + b


# Out[12]:
# array([[ 0,  2,  4],
#        [ 3,  5,  7],
#        [ 6,  8, 10]])


a - b


# array([[0, 0, 0],
#        [3, 3, 3],
#        [6, 6, 6]])


a * b


# Out[11]:
# array([[ 0,  1,  4],
#        [ 0,  4, 10],
#        [ 0,  7, 16]])


a < 5


# Out[12]:
# array([[ True,  True,  True],
#        [ True,  True, False],
#        [False, False, False]])


a ** 2


# Out[13]:
# array([[ 0,  1,  4],
#        [ 9, 16, 25],
#        [36, 49, 64]], dtype=int32)


a += 3
a


# Out[17]:
# array([[ 3,  4,  5],
#        [ 6,  7,  8],
#        [ 9, 10, 11]])

#######################################
# 矩阵内置操作
#######################################
from numpy import array as matrix, arange


a = arange(9).reshape(3,3)
a


# Out[10]:
# array([[0, 1, 2],
#        [3, 4, 5],
#        [6, 7, 8]])


a.max()


# Out[23]: 8


a.min()


# Out[24]: 0


a.sum()


# Out[25]: 36

#######################################
# 矩阵索引、拆分、遍历
#######################################
from numpy import array as matrix, arange


a = arange(25).reshape(5,5)
a


# Out[9]:
# array([[ 0,  1,  2,  3,  4],
#        [ 5,  6,  7,  8,  9],
#        [10, 11, 12, 13, 14],
#        [15, 16, 17, 18, 19],
#        [20, 21, 22, 23, 24]])


a[2,3]      # 取第3行第4列的元素


# Out[3]: 13


a[0:3,3]    # 取第1到3行第4列的元素


# Out[4]: array([ 3,  8, 13])


a[:,2]      # 取所有第二列元素


# Out[7]: array([ 2,  7, 12, 17, 22])


a[0:3,:]    # 取第1到3行的所有列


# Out[8]:
# array([[ 0,  1,  2,  3,  4],
#        [ 5,  6,  7,  8,  9],
#        [10, 11, 12, 13, 14]])


a[-1]   # 取最后一行


# Out[10]: array([20, 21, 22, 23, 24])


for row in a:   # 逐行迭代
    print(row)


# [0 1 2 3 4]
# [5 6 7 8 9]
# [10 11 12 13 14]
# [15 16 17 18 19]
# [20 21 22 23 24]


for element in a.flat:  # 逐元素迭代，从左到右，从上到下
    print(element)


# 0
# 1
# 2
# 3
# ...

#######################################
# 改变矩阵
#######################################
from numpy import array as matrix, arange


b = arange(20).reshape(5,4)


b


# Out[18]:
# array([[ 0,  1,  2,  3],
#        [ 4,  5,  6,  7],
#        [ 8,  9, 10, 11],
#        [12, 13, 14, 15],
#        [16, 17, 18, 19]])


b.ravel()


# Out[16]:
# array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
#        17, 18, 19])


b.reshape(4,5)


# Out[17]:
# array([[ 0,  1,  2,  3,  4],
#        [ 5,  6,  7,  8,  9],
#        [10, 11, 12, 13, 14],
#        [15, 16, 17, 18, 19]])


b.T     # reshape 方法不改变原矩阵的值，所以需要使用  .T  来获取改变后的值


# Out[19]:
# array([[ 0,  4,  8, 12, 16],
#        [ 1,  5,  9, 13, 17],
#        [ 2,  6, 10, 14, 18],
#        [ 3,  7, 11, 15, 19]])

#######################################
# 合并矩阵
#######################################
from numpy import array as matrix,newaxis
import numpy as np


d1 = np.floor(10*np.random.random((2,2)))
d2 = np.floor(10*np.random.random((2,2)))


d1


# Out[7]:
# array([[1., 0.],
#        [9., 7.]])


d2


# Out[9]:
# array([[0., 0.],
#        [8., 9.]])


np.vstack((d1,d2))  # 按列合并


# Out[10]:
# array([[1., 0.],
#        [9., 7.],
#        [0., 0.],
#        [8., 9.]])


np.hstack((d1,d2))  # 按行合并


# Out[11]:
# array([[1., 0., 0., 0.],
#        [9., 7., 8., 9.]])


np.column_stack((d1,d2)) # 按列合并


# Out[13]:
# array([[1., 0., 0., 0.],
#        [9., 7., 8., 9.]])


c1 = np.array([11,12])
c2 = np.array([21,22])


np.column_stack((c1,c2))


# Out[14]:
# array([[11, 21],
#        [12, 22]])


c1[:,newaxis]   # 添加一个“空”列


# Out[18]:
# array([[11],
#        [12]])


np.hstack((c1,c2))


# Out[27]: array([11, 12, 21, 22])


np.hstack((c1[:,newaxis],c2[:,newaxis]))


# Out[28]:
# array([[11, 21],
#        [12, 22]])

参考

1. NumPy官方文档