How to Use NumPy Random choice() in Python?

NumPy random.choice() function in Python is used to return a random sample from a given 1-D array. It creates an array and fills it with random samples. It has four parameters and using these parameters we can manipulate the random samples of an array.

In this article, I will explain how to use the NumPy random.choice() function and using its syntax, parameters, and how to generate random samples of a given 1-D array with examples.

1. Quick Examples of random.choice() Function

Following are quick examples of random.choice()


# Quick examples of random.choice() function

# Example 1: Get the single element from random choice
arr = np.random.choice(7)

# Example 2:  Get an array of uniform random samples 
arr = np.random.choice(5, 5) 

# Example 3: Get an array of non uniform random samples  
arr = np.random.choice(5, 5, p=[0.1, 0, 0.3, 0.6, 0])

# Example 4: Get the random values without replace
arr = np.random.choice(5, 5, replace = False)

# Example 5: Get the Non-random values without replace
arr = np.random.choice(5, 3, replace = False, p=[0.1, 0, 0.3, 0.6, 0])

2. Syntax of random.choice()

Following is the syntax of the NumPy random.choice() function.


# Syntax of random.choice
random.choice(arr, size=None, replace=True, p=None)

2.1 Parameters of random.choice()

Following are the parameters of random.choice() function.

arr – 1-D NumPy array or int. If a ndarray a random sample is generated from its elements.
size – (Optional) The shape of the output. If None, a single random element is returned. If an integer, size a number of random elements are generated. If a tuple of integers, the output will have that shape.
replace – (optional)Whether the random sample is with or without replacement. Default is True, meaning that a value of arr can be selected multiple times.
p – (Optional) The probabilities associated with each entry in the input array. If specified, the probabilities must sum to 1. If None, the function assumes a uniform distribution.

2.2 Return value

It returns an ndarray of random samples.

3. Usage of NumPy random.choice()

The NumPy random.choice() function is a built-in function in the NumPy module package and is used to create a one-dimensional NumPy array of random samples. We know that the NumPy module is a data manipulation library for Python. Some special tools of NumPy operate on arrays of numbers. For example, the manipulation of numeric data is a big task in data analysis and statistics for getting random data samples.

In the below example, np.random.choice(7) to generate a single random element from the numbers 0 to 6 (inclusive). In this case, np.random.choice(7) will randomly select a single integer from the range [0, 1, 2, 3, 4, 5, 6]. The selected random number will be stored in the variable arr, and it will be printed using the print() statement.


# Import numpy 
import numpy as np

# Get the single element from random choice
arr = np.random.choice(7)
print("After getting the random choice:", arr)

Yields below output.

Note that the output will be different each time you run the code because it’s a random choice.

4. Get Uniform Random Samples of NumPy Array

You can use np.random.choice(5, 5) to generate an array of 5 uniform random samples from the integers 0 to 4 (inclusive). In this case, np.random.choice(5, 5) will generate an array of 5 elements, each randomly chosen from the integers [0, 1, 2, 3, 4]. The resulting array arr will contain 5 random integers.


# Get an array of uniform random samples 
arr = np.random.choice(5, 5)
print("After getting an array of uniform random samples:\n",arr)

Yields below output.

5. Get Non-Uniform random samples of NumPy Array

You can also use np.random.choice(5, 5, p=[0.1, 0, 0.3, 0.6, 0]) to generate an array of 5 non-uniform random samples from the integers 0 to 4 (inclusive) with specified probabilities.

In this program, the p parameter specifies the probabilities associated with each element in the input array. The probabilities [0.1, 0, 0.3, 0.6, 0] indicate the likelihood of each element being chosen. In this example, the second and last elements have a probability of 0, so they will never be selected. The third element has a probability of 0.3, and the fourth element has a probability of 0.6, making them more likely to be chosen.


# Get an array of non uniform random samples  
arr = np.random.choice(5, 5, p=[0.1, 0, 0.3, 0.6, 0])
print("After getting an array of non uniform random samples:\n",arr)

# Output:
# After getting an array of non uniform random samples:
#  [2 3 2 2 2]

6. Get the Uniform Random Sample without Replacement

Create a uniform random sample from arange(5) of size 5 without replacement. which means the selected elements may be repeated, as you can see in the above output few elements are repeated in the randomly selected array. Whereas if replace=False then the elements will not repeat in the randomly selected array.


# Get the random values without replace
arr = np.random.choice(5, 5, replace = False)
print("After getting random values without replace:\n",arr)

# Output:
# After getting random values without replace:
#  [4 2 3 0 1]

7. Get the Non-Uniform Random Sample without Replacement

Create a non-uniform random sample from arange(5) of size 3 without replacement. For that, pass p parameter of the same size as the given array and set replace=False into this function, it will return Non-repeated and Non-uniform random samples of the given array.


# Get the Non-random values without replace
arr = np.random.choice(5, 3, replace = False, p=[0.1, 0, 0.3, 0.6, 0])
print("After getting non random values without replace:\n",arr)

# Output:
# After getting non random values without replace:
#  [3 2 0]

8. Get the Graphical presentation of Random Values

Let’s plot the graph of the random values using the matplotlib library.


import matplotlib.pyplot as plt
# Using choice() method
arr = np.random.choice(7, 300)
count, bins, ignored = plt.hist(arr, 25, density=True)
plt.show()

Frequently Asked Questions

What is the purpose of numpy.random.choice()?

numpy.random.choice() is used to generate random samples from a specified 1-D array-like object. It can be used to randomly select elements from an array, generate random integers, or perform random sampling with or without replacement.

How do I generate a random integer between a specific range using numpy.random.choice()?

You can generate a random integer between a specific range using numpy.random.choice() by providing an array-like object containing the range of integers. For example, np.random.choice(7) will generate a random integer between 0 and 6.

How do I generate random samples without replacement using numpy.random.choice()?

To generate random samples without replacement, set the replace parameter to False. For example, np.random.choice(5, 3, replace=False) will generate 3 random samples without replacement from the integers 0 to 4.

Can numpy.random.choice() be used with non-integer data types?

numpy.random.choice() can be used with non-integer data types. It works with any array-like object, including arrays of floats, strings, or other data types.

Can I generate non-uniform random samples using numpy.random.choice()?

You can generate non-uniform random samples by providing the p parameter, which specifies the probabilities associated with each element in the input array. The function will sample elements based on these probabilities.

Conclusion

In this article, I have explained the NumPy random.choice() function syntax, parameter, and usage of how to get the random samples of 1-D NumPy array with examples.

References

https://np.org/doc/stable/reference/random/generated/np.random.choice.html