Random numbers are used in Monte Carlo methods, stochastic calculus, and more. Real random numbers are difficult to produce, so in practice, we use pseudo-random numbers. Pseudo-random numbers are sufficiently random for most intents and purposes, except for some very exceptional instances, such as very accurate simulations. The random-numbers-associated routines can be located in the NumPy random subpackage.
Note
The core random-number generator is based on the Mersenne Twister algorithm (refer to https://en.wikipedia.org/wiki/Mersenne_twister).
Random numbers can be produced from discrete or continuous distributions. The distribution functions have an optional size argument, which informs NumPy how many numbers are to be created. You can specify either an integer or a tuple as the size. This will lead to an array of appropriate shapes filled with random numbers. Discrete distributions include the geometric, hypergeometric, and binomial distributions. Continuous distributions...