JIYIK CN >

In Python, we can easily generate random numbers using the Random and NumPy libraries.

Selecting random elements from a list or array based on the possible outcomes of the elements is called weighted random selection. The selection of elements is determined by assigning a probability to each element present. Sometimes multiple elements are also selected from the list of elements made.

In this tutorial, we will discuss how to generate weighted random selections in Python.

Use random.choices()the function to generate a weighted random selection

Here, Python’s randommodule is used to generate random numbers.

In choices()the function, weighted random selection is performed with replacement. It is also called weighted random sample with replacement. Moreover, in this function, weights play a vital role. Weights define the possible outcomes of selecting each element. There are two types of weights:

1. Relative Weight
2. Cumulative weight

Select elements with relative weights

weightsThe parameters define the relative weights. For each element in the list, the possible outcomes are different. If the possible outcomes for each element have been determined using relative weights, the selection is made based on the relative weights only.

Here is an example:

import random

List = [12, 24, 36, 48, 60, 72, 84]
print(random.choices(List, weights=(30, 40, 50, 60, 70, 80, 90), k=7))

Here each element in the list has its own weight i.e. possible outcome. Also, k in the above example is the number of elements required in the given list.

Output:

[60, 84, 36, 72, 84, 84, 60]

Here, the sum of the weights does not add up to 100 because they are relative weights and not percentages. The number 84 appears 3 times because it has the highest weight among all the weights. So its probability of occurring is the highest.

Select elements with cumulative weights

cum_weightThe parameter is used to define the cumulative weight. The cumulative weight of an element is determined by the weight of the previous element plus the relative weight of the element. For example, the relative weights [10, 20, 30, 40] are equivalent to the cumulative weights [10, 30, 60, 100]

Here is an example:

import random

List = [13, 26, 39, 52, 65]
print(random.choices(List, cum_weights=(10, 30, 60, 100, 150), k=5))

Output:

[65, 65, 39, 13, 52]

Here too, the number 65 appears more often than any other number because it has the highest weight.

Use numpy.random.choice()the function to generate a weighted random selection

To generate random weighted selections, NumPy is typically used when using Python versions lower than 3.6.

Here, numpy.random.choiceis used to determine the probability distribution. In this method, a random element of a one-dimensional array is obtained and choice()a random element of a numpy array is returned using the function.

import numpy as np

List = [500, 600, 700, 800]
sNumbers = np.random.choice(List, 4, p=[0.10, 0.20, 0.30, 0.40])
print(sNumbers)

Here, the probability should be equal to 1. The number 4 represents the size of the list.

Output:

[800 500 600 800]