Generating Random Numbers - Scientific Computing with Python

Numpy provides several functions for generating random numbers. Documentation can be found here.

%pylab inline

import numpy as np
import scipy.stats as stats

Populating the interactive namespace from numpy and matplotlib

Uniform Distrubution¶

np.random.rand samples double precision numbers in the range [0,1] uniformly at random.

np.random.rand()

0.1684372512782496

To generate an array of random numbers:

np.random.rand(2,2)

array([[0.06547767, 0.79321455],
       [0.12412586, 0.17844781]])

np.random.randn samples double precision numbers from a Gaussian (normal) distribution

np.random.randn()

0.38324242962418337

Again, an array of random numbers can be generated:

np.random.randn(2,2)

array([[-1.24202748, -0.19533947],
       [ 0.12978598,  0.83428311]])

np.random.randint samples inegers

np.random.randint(10)

7

to sample and array, size must be specified

np.random.randint(10, size=(2,2))

array([[9, 9],
       [9, 7]])

While numpy provides basic random number generation capabilities, it does not provide utilities for sampling from more complex distribitions.

scipy.stats provides classes for a variety of commonly-used distributions - see the documentation

import scipy.stats as stats

d = stats.poisson(2.0)

You can sample from the distribution using the rvs method

d.rvs(5)

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