Random
ランダムなIndexがほしい時
numpy.random.randint(low, high=None, size=None)
low は含まれるが、highは含まれない。
分布は一様分布
In [49]: np.random.randint(0,100,size=10)
Out[49]: array([91, 64, 71, 65, 94, 58, 53, 14, 49, 52])
In [50]: np.random.randint(0,100,size=(2,3))
Out[50]:
array([[18, 30, 22],
[95, 34, 85]])
In [51]: x = np.random.randint(0,100,size=1000)
一様分布(uniform distribution)
numpy.random.rand(d0, d1, ..., dn)
In [12]: np.random.rand(10)
Out[12]:
array([ 0.88113083, 0.20320019, 0.32004609, 0.37492399, 0.14972709,
0.12742709, 0.75792997, 0.61045003, 0.04126539, 0.89053443])
In [13]: np.random.rand(2,3)
Out[13]:
array([[ 0.9020553 , 0.83688775, 0.28991469],
[ 0.83150826, 0.13279892, 0.12171419]])
In [14]: x = np.random.rand(10000)
正規分布(standard normal distribution)
numpy.random.randn(d0, d1, ..., dn)
In [17]: np.random.randn(10)
Out[17]:
array([-0.82515182, 0.76906557, -0.7280289 , -0.01442234, -1.06237328,
0.39381346, -1.15061995, -0.09937671, 0.66138369, 1.21645816])
In [18]: np.random.randn(2,3)
Out[18]:
array([[ 0.80548817, -1.2232574 , -0.33438786],
[-1.85177797, -1.30872667, -0.28216313]])
In [19]: x = np.random.randn(10000)
In [41]: def N(mean, var, shape):
...: return np.sqrt(var)*np.random.randn(*shape) + mean
...:
...:
In [42]: N(5.0, 10.0, [10])
Out[42]:
array([ 4.86729898, 6.7609417 , 4.26180116, 3.88441367, 5.98194004,
-1.13748272, 5.74417308, 4.0822398 , 8.02914619, 6.88882414])
In [43]: N(5.0, 10.0, [2, 3])
Out[43]:
array([[ 4.97538772, 5.79304735, 0.39961495],
[ 0.38103024, 3.00300634, -0.95105639]])
In [44]: x = N(5.0, 10.0, [10000])
numpy arrayをshuffleしたいとき
numpy.random.shuffle(x)
※引数で渡したデータはshuffleされる
In [55]: A = np.arange(100)
In [56]: np.random.shuffle(A)
In [57]: A
Out[57]:
array([77, 74, 91, 29, 16, 23, 54, 10, 4, 21, 89, 32, 28, 52, 47, 27, 68,
3, 73, 64, 63, 33, 39, 67, 62, 37, 66, 55, 57, 83, 98, 35, 13, 84,
90, 26, 45, 44, 82, 41, 15, 42, 79, 65, 43, 24, 95, 72, 78, 1, 69,
51, 18, 61, 31, 50, 2, 46, 94, 17, 70, 7, 85, 6, 87, 56, 40, 60,
19, 25, 0, 88, 30, 5, 86, 34, 93, 9, 48, 20, 8, 36, 49, 11, 99,
97, 71, 38, 75, 59, 58, 22, 14, 92, 76, 80, 53, 96, 12, 81])
多次元のnumpy arrayをshuffleしたときの動作
多次元のnumpy arrayをshuffleした場合は、最初のindex（最初のdimension）に沿ってshuffleされる。
In [19]: A = np.arange(9).reshape(3,3)
In [20]: print A
[[0 1 2]
[3 4 5]
[6 7 8]]
In [21]: np.random.shuffle(A)
In [22]: print A
[[3 4 5]
[0 1 2]
[6 7 8]]
shuffleされたIndexがほしい時
numpy.random.permutation(x)
データをshuffleしてほしくない時は、IndexをShuffleして渡すと良い
In [58]: A = np.arange(100)
In [59]: x = np.random.permutation(100)
In [60]: A[x]
Out[60]:
array([14, 36, 70, 75, 7, 80, 19, 99, 40, 97, 31, 20, 87, 68, 78, 50, 73,
66, 12, 45, 69, 72, 24, 42, 58, 89, 71, 98, 90, 63, 34, 81, 67, 55,
17, 1, 23, 62, 28, 39, 85, 65, 52, 29, 22, 49, 9, 13, 18, 25, 59,
83, 38, 74, 27, 30, 64, 84, 16, 77, 95, 61, 33, 5, 37, 44, 46, 48,
54, 82, 76, 92, 57, 10, 60, 53, 47, 56, 41, 8, 26, 91, 0, 93, 11,
2, 3, 94, 86, 32, 43, 88, 35, 96, 79, 51, 21, 15, 6, 4])
In [61]: A
Out[61]:
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, 25, 26, 27, 28, 29, 30, 31, 32, 33,
34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50,
51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67,
68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84,
85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99])
１００個あるうちからRandomに５個選ぶとき
In [62]: A[x][:5]
Out[62]: array([14, 36, 70, 75, 7])
リストから指定した確率でデータを選びたいとき
numpy.random.choice(a, size=None, replace=True, p=None)
In [98]: Data = ['A','B','C','D','E']
In [99]: np.random.choice(Data, 10, p=[0.1, 0.4, 0.2, 0.1, 0.2])
Out[99]:
array(['D', 'B', 'B', 'D', 'B', 'B', 'E', 'B', 'E', 'D'],
dtype='|S1')
In [100]: x = np.random.choice(Data, 10000, p=[0.1, 0.4, 0.2, 0.1, 0.2])
h = dict((c, (x==c).sum()/float(len(x))) for c in x)
fig, ax = plt.subplots()
ind = np.arange(len(Data))
ax.set_xticklabels(Data)
ax.set_xticks(ind + 0.5)
ax.bar(ind, [h[k] for k in Data])