This tutorial shows how you can use Numpy to generate random numbers in Python. The following is the basic syntax summarizing 3 functions.
1. Integers:
np.random.randint()
2. Normal distribution:
np.random.randn()
3. Uniform distribution:
np.random.rand()
Example 1: Integer
np.random.randint(low, high=None, size=None, dtype=int)
np.random.randint() will return integer numbers. Given that there are quite a few parameters in randint(), it is better to specify low, high, and size in the statement.
import numpy as np
# Without size information, it will just generate a random integer
Array_1 = np.random.randint(low=2, high=90)
print("First Array: \n", Array_1)
# The following has the size of 10 numbers
Array_2 = np.random.randint(low=2, high=90, size=10)
print("Second Array: \n", Array_2)
# The following has the shape mentioned
Array_3 = np.random.randint(low=2,high=90,size=(3,4))
print("Third Array: \n", Array_3)The following is the output showing Array_1, Array_2, and Array_3 from randint() in Numpy.
First Array: 53 Second Array: [56 79 71 15 27 15 88 32 32 14] Third Array: [[67 33 59 38] [29 20 79 24] [25 13 30 76]]
Example 2: Standard Normal Distribution
np.random.randn(d0, d1, …, dn)
np.random.randn() will return data of a standard normal distribution (i.e., mean=0 and variation = 1.).
d0, d1, …, dn denotes the size in each dimension. If no d0, d1, …, dn is given, a single Python float is returned.
import numpy as np
# Without argument, it will return a single floating point number
Array_1 = np.random.randn()
print("First Array: \n", Array_1)
# d0=10, which returns an array with a size of 10
Array_2 = np.random.randn(10)
print("Second Array: \n", Array_3)
# d0=3, d1=4, size of 3x4 array
Array_3 = np.random.randn(3,4)
print("Third Array: \n", Array_3)The following are the outputs of Array_1, Array_2, and Array_3 from randn() in Numpy.
First Array: 0.06768179188386116 Second Array: [-1.39432412e+00 -4.31202971e-01 -2.69721318e+00 6.02813442e-01 -3.71431990e-01 1.49630972e+00 5.38674312e-04 2.17072606e-01 2.39233874e-01 -5.64958800e-01] Third Array: [[ 1.17880526 -0.14422117 -0.09097434 0.22683969] [-0.10747543 0.68855861 -0.91183705 2.06881669] [ 1.23845381 -0.87994308 0.41059549 -0.55673435]]
Example 3: Uniform Distribution
np.random.rand(d0, d1, …, dn)
np.random.rand() will return data following uniform distribution (i.e., in the range of [0,1)).
d0, d1, …, dn denotes the size in each dimension. If no d0, d1, …, dn is given, a single Python float is returned.
import numpy as np
# without any argument, it will return a single floating point number in the range of [0, 1)
Array_1 = np.random.rand()
print("First Array: \n", Array_1 )
# d0=5; generate 5 numbers in the range of [0, 1)
Array_2 = np.random.rand(5)
print("Second Array: \n", Array_2)
# d0=3, d1=4, size of 3x4 array; generate 12 numbers in the range of [0, 1)
Array_3 = np.random.rand(3,4)
print("Third Array: \n", Array_3)The following are the outputs of Array_1, Array_2, and Array_3 from rand() in Numpy.
First Array: 0.11727791343839256 Second Array: [0.71103727 0.63259931 0.54372395 0.59553258 0.62353549] Third Array: [[0.18877663 0.26831789 0.2800152 0.09025151] [0.63440239 0.09823381 0.2045029 0.94778125] [0.23219263 0.10362962 0.86029194 0.89346368]]
Further Reading
The following tutorial is specifically about generating a random sample of normal distribution in Python.