Numpy-04常用random随机函数汇总

Numpy常用random随机函数汇总

函数名	说明
seed([seed])	设定随机种子，这样每次生成的随机数会相同
rand(d0, d1, …, dn)	返回数据在[0,1）之间，具有均匀分布
randn(d0, d1, …, dn)	返回数据具有标准正态分布（均值0， 方差1）
randint(low[, high, size, dtype])	生成随机整数，包含low，不包含high
random([size])	生成[0.0, 1.0)的随机数
choice(a [, size, replace, p])	a是一维数组，从它里面生成随机结果
shuffle(x)	把一个数组x进行随机排列
permutation(x)	把一个数组x进行随机排列，或者数字的全排列
normal([loc, scale, size])	按照平均值loc和方差scale生成高斯分布的数字
uniform([low, high, size])	在[low，high）之间生成均匀分布的数字

---

In [1]:

import numpy as np
np.random.seed(666)

1.rand(d0, d1, …, dn)

返回数据在[0, 1)之间，具有均匀分布

In [2]:

np.random.rand(5)

Out[2]:

array([0.70043712, 0.84418664, 0.67651434, 0.72785806, 0.95145796])

In [3]:

np.random.rand(3, 4)

Out[3]:

array([[0.0127032 , 0.4135877 , 0.04881279, 0.09992856],
       [0.50806631, 0.20024754, 0.74415417, 0.192892  ],
       [0.70084475, 0.29322811, 0.77447945, 0.00510884]])

In [4]:

np.random.rand(2, 3, 4)

Out[4]:

array([[[0.11285765, 0.11095367, 0.24766823, 0.0232363 ],
        [0.72732115, 0.34003494, 0.19750316, 0.90917959],
        [0.97834699, 0.53280254, 0.25913185, 0.58381262]],

       [[0.32569065, 0.88889931, 0.62640453, 0.81887369],
        [0.54734542, 0.41671201, 0.74304719, 0.36959638],
        [0.07516654, 0.77519298, 0.21940924, 0.07934213]]])

2.randn(d0, d1, …, dn)

返回数据具有标准正态分布（均值0，方差1）

In [5]:

np.random.randn(5)

Out[5]:

array([-1.20990266, -0.04618272, -0.44118244,  0.46953431,  0.44325817])

In [6]:

np.random.randn(3, 4)

Out[6]:

array([[-1.66738875, -1.81731749, -1.39753916,  0.78392691],
       [-0.29129965,  0.67049043,  0.706931  ,  1.42965241],
       [-0.41407013, -1.32672274, -0.14880188,  0.34771289]])

In [7]:

np.random.randn(2, 3, 4)

Out[7]:

array([[[ 0.61030791, -1.17532603,  0.82985368, -0.30236752],
        [-0.04327047,  0.06706965, -1.59102817,  0.01705112],
        [-1.87296591, -0.96457904, -0.00420389,  0.47495047]],

       [[-0.05421452,  0.89181355,  0.96866859,  0.6307865 ],
        [-0.89051986,  0.08227022, -0.07594056,  0.42969347],
        [ 0.11579967, -0.54443241,  0.02835341,  1.34408655]]])

3.randint(low[, high, size, dtype])

生成随机整数，包含low，不包含high。如果high不指定，则从[0, low)中生成数字

In [8]:

np.random.randint(3)

Out[8]:

2

In [12]:

np.random.randint(1, 10)

Out[12]:

9

In [13]:

np.random.randint(10, 30, size=(5,))

Out[13]:

array([18, 26, 19, 14, 21])

In [14]:

np.random.randint(10, 30, size=(2, 3, 4))

Out[14]:

array([[[29, 25, 19, 21],
        [28, 12, 13, 19],
        [27, 27, 18, 27]],

       [[16, 24, 16, 19],
        [21, 20, 19, 14],
        [22, 26, 29, 20]]])

4.random([size])

生成[0.0, 1.0)的随机数

In [15]:

np.random.random(5)

Out[15]:

array([0.67940357, 0.65592832, 0.0039444 , 0.06654134, 0.00112109])

In [16]:

np.random.random(size=(3, 4))

Out[16]:

array([[0.66608382, 0.38565116, 0.09405827, 0.45856757],
       [0.64434173, 0.59499774, 0.79060307, 0.79996907],
       [0.67969792, 0.43875185, 0.26235889, 0.23652188]])

In [17]:

np.random.random(size=(2, 3, 4))

Out[17]:

array([[[0.83900208, 0.36874334, 0.61918838, 0.46656433],
        [0.49250063, 0.71862211, 0.65415881, 0.9665017 ],
        [0.38957233, 0.97017219, 0.36057961, 0.56184234]],

       [[0.03133558, 0.30480028, 0.07269465, 0.46721993],
        [0.41345069, 0.42228271, 0.79491031, 0.05651855],
        [0.89718201, 0.31869638, 0.36398678, 0.70548804]]])

5.choice(a[, size, replace, p])

a是一维数组，从它里面生成的随机结果

In [18]:

# 这时候，a是数字，则从range（5）中生成，size为3
np.random.choice(5, 3)

Out[18]:

array([3, 2, 4])

In [19]:

np.random.choice(5, (2, 3))

Out[19]:

array([[0, 2, 1],
       [4, 2, 3]])

In [20]:

# 这时候，a是数组，从里面随机取出数组
np.random.choice([2, 3, 6, 7, 9], 3)

Out[20]:

array([9, 2, 6])

In [21]:

np.random.choice([2, 3, 6, 7, 9], (2, 3))

Out[21]:

array([[9, 6, 2],
       [9, 2, 3]])

6.shuffle(x)

把一个数组x进行随机排列

In [22]:

a = np.arange(10)
np.random.shuffle(a)
a

Out[22]:

array([1, 9, 3, 2, 4, 6, 0, 5, 7, 8])

In [25]:

a = np.arange(20).reshape(4, 5)
a

Out[25]:

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

In [26]:

# 如果数组是多维的，则只会在第一维度打散数据  只把行数打散
np.random.shuffle(a)
a

Out[26]:

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

7.permutation(x)

把一个数组x进行随机排列，或者数字的全排列

In [27]:

# 这时候，生成range（10）的随机排列
np.random.permutation(10)

Out[27]:

array([2, 5, 8, 1, 9, 0, 6, 4, 3, 7])

In [29]:

# 这时候，在第一维度进行打散
arr = np.arange(9).reshape((3, 3))
arr

Out[29]:

array([[0, 1, 2],
       [3, 4, 5],
       [6, 7, 8]])

In [30]:

# 注意，这里不会更改原来的arr，会返回一个新的copy
np.random.permutation(arr)

Out[30]:

array([[0, 1, 2],
       [6, 7, 8],
       [3, 4, 5]])

8.normal([loc, scale, size])

按照平均值loc和方差scale生成高斯分布的数字

In [31]:

np.random.normal(1, 10, 10)

Out[31]:

array([ 10.80291099,   9.2194    , -13.45464412,   5.92066144,
         2.55578918,   4.70191278,   1.25552781,  19.16839105,
       -11.2572319 , -13.4167598 ])

In [32]:

np.random.normal(1, 10, (3, 4))

Out[32]:

array([[ 13.53763269,  13.41439272,   7.04674961,   1.11736277],
       [ -7.16656663,  -9.45062631,   5.06354375,  -7.57895331],
       [  8.05614584,   2.69688786,  23.53394139, -13.75199322]])

9.uniform([low, high, size])

在[low, high)之间生成均匀分布的数字

In [33]:

np.random.uniform(1, 10, 10)

Out[33]:

array([3.1972772 , 6.84872322, 4.33722237, 8.13376077, 8.64726595,
       8.12580587, 2.93863321, 7.52063524, 7.98059166, 6.48754099])

In [34]:

np.random.uniform(1, 10, (3, 4))

Out[34]:

array([[3.5182339 , 7.30229967, 5.12969041, 9.02342506],
       [6.99628706, 5.64623389, 7.32123531, 6.77599624],
       [9.30139114, 2.38426176, 9.42378508, 4.75697765]])

实例：对数组加入随机噪声

In [35]:

import matplotlib.pyplot as plt

In [36]:

# 绘制sin曲线
x = np.linspace(-10, 10, 100)    # 指定最小数、最大数，在它们中间生成100个点
y = np.sin(x)
plt.plot(x, y)
plt.show()

In [37]:

# 加入噪声
x = np.linspace(-10, 10, 100)
y = np.sin(x) + np.random.rand(len(x))    # len(x)生成的维度和x，y是一样的
plt.plot(x, y)
plt.show()