| 函数 | 说明 |
|---|---|
| rand(d0,d1,..,dn) | 根据d0‐dn创建随机数数组，浮点数， [0,1)，均匀分布 |
| randn(d0,dn) | 根据d0‐dn创建随机数数组，标准正态分布 |
| randint(low[,high,shape]) | 根据shape创建随机整数或整数数组，范围是[low,high) |
| seed(s) | 随机数种子， s是给定的种子值 |
np.random.rand
import numpy as np a = np.random.rand(3,4,5) a Out[3]: array([[[0.28576737,0.96566496,0.59411491,0.47805199,0.97454449],[0.15970049,0.35184063,0.66815684,0.13571458,0.41168113],[0.66737322,0.91583297,0.68033204,0.49083857,0.33549182],[0.52797439,0.23526146,0.39731129,0.26576975,0.26846021]],[[0.46860445,0.84988491,0.92614786,0.76410349,0.00283208],[0.88036955,0.01402271,0.59294569,0.14080713,0.72076521],[0.0537956,0.08118672,0.59281986,0.60544876,0.77931621],[0.41678215,0.24321042,0.25167563,0.94738625,0.86642919]],[[0.36137271,0.21672667,0.85449629,0.51065516,0.16990425],[0.97507815,0.78870518,0.36101021,0.56538782,0.56392004],[0.93777677,0.73199966,0.97342172,0.42147127,0.73654324],[0.83139234,0.00221262,0.51822612,0.60964223,0.83029954]]])
np.random.randn
b = np.random.randn(3,5) b Out[5]: array([[[ 0.09170952,-0.36083675,-0.18189783,-0.52370155,-0.61183783],[ 1.05285606,-0.82944771,-0.93438396,0.32229904,-0.85316565],[ 1.41103666,-0.32534111,-0.02202953,1.02101228,1.59756695],[-0.33896372,0.42234042,0.14297587,-0.70335248,0.29436318]],[[ 0.73454216,0.35412624,-1.76199508,1.79502353,1.05694614],[-0.42403323,-0.36551581,0.54033378,-0.04914723,1.15092556],[ 0.48814148,1.09265266,0.65504441,-1.04280834,0.70437122],[ 2.92946803,-1.73066859,-0.30184912,1.04918753,-1.58460681]],[[ 1.24923498,-0.65467868,-1.30427044,1.49415265,0.87520623],[-0.26425316,-0.89014489,0.98409579,1.13291179,-0.91343016],[-0.71570644,0.81026219,-0.00906133,0.90806035,-0.914998 ],[ 0.22115875,-0.81820313,0.66359573,-0.1490853,0.75663096]]])
np.random.randint
c = np.random.randint(100,200,(3,4)) c Out[9]: array([[104,140,161,193],[134,147,126,120],[117,141,162,137]])
numpy.random.randint的详细用法 - python
函数的作用是，返回一个随机整型数，范围从低（包括）到高（不包括），即[low,high)。如果没有写参数high的值，则返回[0,low)的值。
numpy.random.randint(low,high=None,size=None,dtype='l')
参数如下：
| 参数 | 描述 |
|---|---|
| low: int | 生成的数值最低要大于等于low。 （hign = None时，生成的数值要在[0,low)区间内） |
| high: int (可选) | 如果使用这个值，则生成的数值在[low,high)区间。 |
| size: int or tuple of ints(可选) | 输出随机数的尺寸，比如size=(m * n* k)则输出同规模即m * n* k个随机数。默认是None的，仅仅返回满足要求的单一随机数。 |
| dtype: dtype(可选)： | 想要输出的格式。如int64、int等等 |
输出：
例子
>>> np.random.randint(2,size=10)
array([1,1,0])
>>> np.random.randint(1,size=10)
array([0,0])>>> np.random.randint(5,size=(2,4))
array([[4,2,1],
[3,0]])>>>np.random.randint(2,high=10,3))
array([[6,8,7],
[2,5,2]])
np.random.seed
随机种子生成器，使下一次生成的随机数为由种子数决定的“特定”的随机数，如果seed中参数为空，则生成的随机数“完全”随机。参考和文档。
np.random.seed(10) np.random.randint(100,4)) Out[11]: array([[109,115,164,128],[189,193,129,108],[173,100,136]]) np.random.seed(10) np.random.randint(100,4)) Out[13]: array([[109,136]])
| 函数 | 说明 |
|---|---|
| shuffle(a) | 根据数组a的第1轴（也就是最外层的维度）进行随排列，改变数组x |
| permutation(a) | 根据数组a的第1轴产生一个新的乱序数组，不改变数组x |
| choice(a[,size,replace,p]) | 从一维数组a中以概率p抽取元素，形成size形状新数组replace表示是否可以重用元素，默认为False |
np.random.shuffle
a = np.random.randint(100,4)) a Out[15]: array([[116,111,154,188],[162,133,172,178],[149,151,177]]) np.random.shuffle(a) a Out[17]: array([[116,177],178]]) np.random.shuffle(a) a Out[19]: array([[162,[116,177]])
可以看到，a发生了变化，轴。
np.random.permutation
b = np.random.randint(100,4)) b Out[21]: array([[113,192,186,130],[130,189,112,165],[131,157,136,127]]) np.random.permutation(b) Out[22]: array([[113,127]]) b Out[24]: array([[113,127]])
可以看到，b没有发生改变。
np.random.choice
c = np.random.randint(100,(8,)) c Out[26]: array([123,194,128,174,188,109,115]) np.random.choice(c,2)) Out[27]: array([[111,123],[109,115],[123,128]])#默认可以出现重复值 np.random.choice(c,2),replace=False) Out[28]: array([[188,111],[174,128]])#不允许出现重复值 np.random.choice(c,p=c/np.sum(c)) Out[29]: array([[194,109]])#指定每个值出现的概率
| 函数 | 说明 |
|---|---|
| uniform(low,size) | 产生具有均匀分布的数组,low起始值,high结束值,size形状 |
| normal(loc,scale,size) | 产生具有正态分布的数组,loc均值,scale标准差,size形状 |
| poisson(lam,size) | 产生具有泊松分布的数组,lam随机事件发生率,size形状 |
u = np.random.uniform(0,10,4)) u Out[31]: array([[9.83020867,4.67403279,8.75744495,2.96068699],[1.31291053,8.42817933,6.59036304,5.95439605],[4.36353698,3.56250327,5.87130925,1.49471337]]) n = np.random.normal(10,4)) n Out[33]: array([[ 8.17771928,4.17423265,3.28465058,17.2669643 ],[10.00584724,9.94039808,13.57941572,4.07115727],[ 6.81836048,6.94593078,3.40304302,7.19135792]]) p = np.random.poisson(2.0,4)) p Out[35]: array([[0,[2,3],[4,3]])
数据分析师分析问题第一步，必须明确这是不是一个问题！！！