Numpy 基础学习笔记

Numpy对矩阵的操作更加便捷，前一段时间的使用中，总会因为一些基础的东西模糊不清而出错，所以今天看了一遍莫烦 的numpy 教程。顺便做了笔记供之后查阅。

Numpy 使用

基本使用

列表转numpy矩阵

import numpy as nparray = np.array([[1,2,3],                  [4,5,6]])

输出

[[1 2 3] [4 5 6]]

---

矩阵的信息

print('number of dim:' ,array.ndim)print('shape:',array.shape)print('size:',array.size)

输出

number of dim: 2shape: (2, 3)size: 6

---

指定元素数据的类型

a = np.array([1,2,3],dtype=np.int)print(a.dtype)b = np.array([1,2,3],dtype=np.int64)print(b.dtype)

输出

int32int64

---

创建一些特殊的矩阵

c = np.zeros((2,4))print(c)d = np.ones((3,5))print(d)e = np.arange(0,20,2).reshape((2,5))print(e)f = np.linspace(1,10,4)print(f)

输出

[[ 0.  0.  0.  0.] [ 0.  0.  0.  0.]][[ 1.  1.  1.  1.  1.] [ 1.  1.  1.  1.  1.] [ 1.  1.  1.  1.  1.]][[ 0  2  4  6  8] [10 12 14 16 18]][  1.   4.   7.  10.]

---

矩阵简单运算

a = np.array([10,20,30,40])b = np.arange(4)print (a,b)print ('a-b:',a-b)print ('a+b:',a+b)print ('b^2:',b**2)print ('b<3:',b<3)

输出

[10 20 30 40] [0 1 2 3]a-b: [10 19 28 37]a+b: [10 21 32 43]b^2: [0 1 4 9]b<3: [ True  True  True False]

---

两种乘法运算

a = np.array([[1,1],             [0,1]])b = np.arange(4).reshape((2,2))print(a)print(b)print('a*b: \n',a*b)print('a.dot(b): \n',a.dot(b))

输出

[[1 1] [0 1]][[0 1] [2 3]]a*b: [[0 1] [0 3]]a.dot(b): [[2 4] [2 3]]

---

求和求大小

a = np.random.random((2,4))print(a)print(np.sum(a))print(np.max(a))print(np.min(a))print(np.sum(a,axis = 1))   # 行求和

输出

[[ 0.16971606  0.67080715  0.57505605  0.59053659] [ 0.94514567  0.24781893  0.28458866  0.82784495]]sum:  4.31151405263max:  0.945145668017min:  0.169716059936行求和： [ 2.00611585  2.3053982 ]

---

其他运算

A = np.arange(2,14).reshape((3,4))print(A)print('min index: ',np.argmin(A))print('max index: ',np.argmax(A))print('mean: ',np.mean(A))print('median: ',np.median(A))print('cumsum: ',np.cumsum(A))  # 累加求和print('diff: \n',np.diff(A))    # 相邻求差

输出

[[ 2  3  4  5] [ 6  7  8  9] [10 11 12 13]]min index:  0max index:  11mean:  7.5median:  7.5cumsum:  [ 2  5  9 14 20 27 35 44 54 65 77 90]diff: [[1 1 1] [1 1 1] [1 1 1]]

---

排序 转置 截取

B = np.random.random((2,4))print('sort:\n',np.sort(B))    #默认行排序print('A.T:\n',A.T)print('A.transpose():\n',A.transpose())print('np.clip(A,4,9):\n',np.clip(A,4,9))

输出

sort: [[ 0.00157543  0.15291127  0.65574215  0.73835591] [ 0.08381171  0.1036112   0.25088152  0.48191754]]A.T: [[ 2  6 10] [ 3  7 11] [ 4  8 12] [ 5  9 13]]A.transpose(): [[ 2  6 10] [ 3  7 11] [ 4  8 12] [ 5  9 13]]np.clip(A,4,9): [[4 4 4 5] [6 7 8 9] [9 9 9 9]]

数组索引操作

import numpy as npA = np.arange(3,15)print('A:\n',A)print('a[0]:',A[0])B = np.arange(3,15).reshape((3,4))print('B:\n',B)print('B[1]:',B[1])print('B[1][2]:',B[1][2])print('B[1,2]:',B[1,2])   # B[1,2]与B[1][2]等价print('B[1,:]:',B[1,:])print('B[1,1:3]:',B[1,1:3])  print('-------------------')for row in B:        # 迭代行，迭代列可以转置后迭代行    print(row)

输出

A: [ 3  4  5  6  7  8  9 10 11 12 13 14]a[0]: 3B: [[ 3  4  5  6] [ 7  8  9 10] [11 12 13 14]]B[1]: [ 7  8  9 10]B[1][2]: 9B[1,2]: 9B[1,:]: [ 7  8  9 10]B[1,1:3]: [8 9]-------------------[3 4 5 6][ 7  8  9 10][11 12 13 14]

---

矩阵合并

主要包含vstack,hstack,concatenate三个函数

import numpy as npA = np.array([1,1,1])B = np.array([2,2,2])C = np.vstack((A,B))D = np.hstack((A,B))print('A:\n',A)print('B:\n',B)print('C:\n',C)print('D:\n',D)"""---------------"""print('A:',A)print('A.T:',A.T)print('A.shape:',A.shape)print('A.T.shape:',A.T.shape)# 无法通过转置把序列别换维度bb = A[np.newaxis,:]print('bb: \n',bb)print('bb.shape:',np.shape(bb))cc = A[:,np.newaxis]print('cc: \n',cc)print('cc.shape:',np.shape(cc))"""-----------------"""A = np.array([[1,1,1,1]])B = np.array([[2,2,2,2]])E = np.concatenate((A,B),axis=0)print(E)

输出

A: [1 1 1]B: [2 2 2]C: [[1 1 1] [2 2 2]]D: [1 1 1 2 2 2]A: [1 1 1]A.T: [1 1 1]A.shape: (3,)A.T.shape: (3,)bb: [[1 1 1]]bb.shape: (1, 3)cc: [[1] [1] [1]]cc.shape: (3, 1)[[1 1 1 1] [2 2 2 2]]

分割

import numpy as npA = np.arange(12).reshape((3,4))print('A:\n',A)print('------------------')print(np.split(A,2,axis=1))print('------------------')print(np.vsplit(A,3))print('------------------')print(np.hsplit(A,2))split(A,2))

输出

A: [[ 0  1  2  3] [ 4  5  6  7] [ 8  9 10 11]]------------------[array([[0, 1],       [4, 5],       [8, 9]]), array([[ 2,  3],       [ 6,  7],       [10, 11]])]------------------[array([[0, 1, 2, 3]]), array([[4, 5, 6, 7]]), array([[ 8,  9, 10, 11]])]------------------[array([[0, 1],       [4, 5],       [8, 9]]), array([[ 2,  3],       [ 6,  7],       [10, 11]])]

数组拷贝

简单的=，是把两个变量指向同一个地址，变还其中的一个内容会影响另一个矩阵的内容。

import numpy as npa = np.array([0,1,2,3])b = aa[0] = 100print('b:\n',b)print('b is a?',b is a)print('----------------')c = a.copy()print('c before a change:',c)a[1] = 111print('c after a change:',c)

输出

b: [100   1   2   3]b is a? True----------------c before a change: [100   1   2   3]c after a change: [100   1   2   3]