Numpy array matrix operation

Learning goals ： Study numpy Array （ matrix ） establish , Information view , Transformation

for example ：

• One day master numpy Array operation

---

Learning content ：

for example ：

1. numpy What libraries are used for
2. numpy Library array （ matrix ） establish
3. Array （ matrix ） Information view
4. Master arrays （ matrix ） Transformation
5. For detailed study suggestions, go to see the official api

---

Learning time ：

Tips ：2022/7/24

Learning output ：

**

1numpy What libraries are used for

**

numpy Library for Scientific Computing , It is the basic library of other scientific computing libraries , It is widely used in open source projects , Such as ：Pandas、Seaborn、Matplotlib、scikit-learn etc. .

Here is the learning matrix （ Mathematical name ）, Computer related majors are called arrays （ data structure ）, It's called matrix here ,out For output results

*2 Matrix creation *

2.1 Use python data structure list,tuple To a matrix
1 That's ok 4 Column , And 2 That's ok 4 Column results

#python data structure conversion array
import numpy as np
arry1=np.array([1,2,3,4])
arry2=np.array((5,6,7,8))
arry3=np.array([[9,10,11,12]
               ,[13,14,15,16]])
print(arry1,arry2,arry3)
out:
[1 2 3 4] [5 6 7 8] [[ 9 10 11 12]
 [13 14 15 16]]

2.2 Use numpy Function quick creation
arange()0 To 10 The interval is 2, a line 5 Column

np.arange(0,10,2)
out:
array([0, 2, 4, 6, 8])

3 That's ok 3 Column 0 Matrix

np.zeros([3,3])
out:
array([[0., 0., 0.],
       [0., 0., 0.],
       [0., 0., 0.]])

3 That's ok 3 Column Unit matrix

np.ones([3,3])
out:
array([[1., 1., 1.],
       [1., 1., 1.],
       [1., 1., 1.]])

linspace()0 To 2 Divide into 10 branch 1 That's ok 10 Column Row vector
logspace() hold 0 To 2 Divide into 5-1 branch , The tolerance is （0-2）/(5-1)）=0.5, With 10 Bottom ,10^0 ,10^0.5 ,10^1 ,10^2 A sequence of equal differences

#create a array have float
#create a array have float 10**0-10**2
print(np.linspace(0,2,10),'\n',np.logspace(0,2,5))
out:
[0.         0.22222222 0.44444444 0.66666667 0.88888889 1.11111111
 1.33333333 1.55555556 1.77777778 2.        ] 
 [  1.           3.16227766  10.          31.6227766  100.        ]

5 That's ok 5 Column positive diagonal matrix

np.diag([1,1,1,1,1])
out:
array([[1, 0, 0, 0, 0],
       [0, 1, 0, 0, 0],
       [0, 0, 1, 0, 0],
       [0, 0, 0, 1, 0],
       [0, 0, 0, 0, 1]])

6 ride 3 That's ok 3 Column unit matrix

np.full((3,3),6)
out:
array([[6, 6, 6],
       [6, 6, 6],
       [6, 6, 6]])

randint()100 To 200 The random number of , Generate 2 That's ok 4 Column matrices

# random a 2*4 rarry
np.random.randint(100,200,size=(2,4))
out:
array([[153, 199, 185, 113],
       [150, 130, 116, 142]])

random.rand(3,2)0 To 1 Random generation 3 That's ok 2 Column matrices

#generate 3 rwo 2 column arry
arr1,arr2=np.random.rand(3,2),np.random.rand(5)
# generate 5 number from 0-1
print('arr1',arr1,'\n arr2',arr2)
out:
arr1 [[0.29692384 0.17947901]
 [0.02604587 0.7050282 ]
 [0.18891757 0.93545309]] 
 arr2 [0.1219717  0.53417921 0.97106664 0.25865028 0.817187  ]

**

3 View matrix information

**
ndim The rank of a matrix ,shape matrix （ That's ok , Column ） dtype data type
astype() Data type transformation

wep1=np.array([[0,1,2,3]
              ,[1,3,5,7]
              ,[2,4,6,8]])
print('array ndim :',wep1.ndim)
print('array shape:',wep1.shape)
print('size:',wep1.size)
# set array shape to 4,3
wep1.shape=4,3
print('dataType:',wep1.dtype)
print('new array',wep1)
print('set data type->',wep1.astype(np.str).dtype)
out:
array ndim : 2
array shape: (3, 4)
size: 12
dataType: int32
new array [[0 1 2]
 [3 1 3]
 [5 7 2]
 [4 6 8]]
set data type-> <U11

**

4 Matrix row column transformation

**
-1 Automatic calculation , The row vector becomes 3 That's ok 2 Column ,2 That's ok 3 Column matrices

# change arry dimension ,-1 auto calcuate
arr3=np.arange(6)
arr3.reshape(3,-1)
arr3.reshape(-1,3)
out:
array([[0, 1, 2],
       [3, 4, 5]])

3 That's ok 2 The column matrix becomes a row vector

# make array to 1 dimension 
wep2=np.arange(6).reshape(3,-1)
print(' result 1\n',wep2.ravel(),'\n result 2\n',wep2.flatten())
out:
 result 1
 [0 1 2 3 4 5] 
 result 2
 [0 1 2 3 4 5]

2 individual 2 That's ok 4 The column matrix is arranged horizontally , Merge vertically into one

# array merge hstack=transverse=axis=1;vstack=longitudinal=axis=0
wep3,wep4=np.arange(8).reshape(2,-1),np.array(((7,7,7,7),(8,8,8,8)))
print(' transverse \n',np.hstack((wep3,wep4)),'\n longitudinal \n',np.vstack((wep3,wep4)))
p1,p2=np.concatenate((wep3,wep4),axis=1),np.concatenate((wep3,wep4),axis=0)
print('p1',p1,'\n p2 \n',p2)
out:
 transverse 
 [[0 1 2 3 7 7 7 7]
 [4 5 6 7 8 8 8 8]] 
 longitudinal 
 [[0 1 2 3]
 [4 5 6 7]
 [7 7 7 7]
 [8 8 8 8]]
p1 [[0 1 2 3 7 7 7 7]
 [4 5 6 7 8 8 8 8]] 
 p2 
 [[0 1 2 3]
 [4 5 6 7]
 [7 7 7 7]
 [8 8 8 8]]

4 That's ok 2 The column matrix is split horizontally and vertically

arr4=np.ones([2,4]).reshape(4,-1)
print(np.vsplit(arr4,2),'\n',np.hsplit(arr4,1))
print('transverse 1 spilt 1',np.split(arr4,1,axis=1),
'\nlongitudinal 1 spilt 4\n',np.split(arr4,4,axis=0))
out:
[array([[1., 1.],
       [1., 1.]]), array([[1., 1.],
       [1., 1.]])] 
 [array([[1., 1.],
       [1., 1.],
       [1., 1.],
       [1., 1.]])]
transverse 1 spilt 1 [array([[1., 1.],
       [1., 1.],
       [1., 1.],
       [1., 1.]])] 
longitudinal  1 spilt 4
 [array([[1., 1.]]), array([[1., 1.]]), array([[1., 1.]]), array([[1., 1.]])]

Matrix transposition

# array transpose row to column
arr5=np.arange(6).reshape(3,2)
print('arr5\n',arr5)
print(arr5.transpose((1,0)))
print(arr5.T)
out:
arr5
 [[0 1]
 [2 3]
 [4 5]]
[[0 2 4]
 [1 3 5]]
[[0 2 4]
 [1 3 5]]