如何使用Numpy生成接近1的随机数

Question

我想生成一个随机数接近1的numpy数组。有没有一种快速的方法可以让我从1（例如1e-5）设置所需的邻域？

Answer 1

Check the numpy.random module:

For example, normally distributed numbers with mean of 1.0 and standard deviation of 0.002:

>>> numpy.random.normal(1, 0.002, (5,))
array([1.00246167, 0.99722898, 0.99793482, 1.00100399, 1.00004228])

Using uniform distribution:

>>> numpy.random.uniform(1-1e-5, 1+1e-5, (5,))
array([1.00000668, 1.00000037, 0.99999398, 0.99999736, 1.00000645])

Answer 2

例如，如果您想要在[1 - 1e-5, 1 + 1e-5)范围内的1000个统一随机数：

nums = np.random.uniform(low=1-1e-5, high=1+1e-5, size=1000)

Answer 3

Use a uniform distribution:

>>> window = 1e-5
>>> np.random.uniform(low=1-window, high=1+window, size=10)
array([ 1.00000539,  0.99999055,  1.00000759,  0.99999228,  1.00000737,
        1.00000557,  1.00000522,  1.00000375,  1.00000054,  0.99999047])