Random Permutations in NumPy

Welcome to The Coding College, your trusted platform for mastering programming concepts! In this article, we’ll explore random permutations in NumPy, a powerful feature used for shuffling data, random sampling, and more.

What is a Random Permutation?

A random permutation refers to rearranging the elements of an array or sequence in random order. This is particularly useful in:

  • Shuffling datasets for machine learning.
  • Creating randomized experiments.
  • Games and simulations.

NumPy provides tools to generate random permutations efficiently using the numpy.random module.

The numpy.random.permutation() Function

The numpy.random.permutation() function generates a randomly permuted sequence or array.

Syntax

numpy.random.permutation(x)
  • x: Can be an integer or an array-like object.

1. Permuting a Sequence

Example: Permute an Array

import numpy as np

arr = np.array([1, 2, 3, 4, 5])
shuffled = np.random.permutation(arr)
print(shuffled)

Output:

[4 1 5 2 3]  # Example output

The original array remains unchanged:

print(arr)  # Output: [1 2 3 4 5]

2. Generating Permutations of Numbers

If an integer is provided, np.random.permutation() returns a shuffled range of numbers.

Example: Generate Permutation of a Range

shuffled_range = np.random.permutation(10)
print(shuffled_range)

Output:

[4 8 2 5 1 6 7 9 0 3]  # Example output

3. Shuffling Multidimensional Arrays

The numpy.random.permutation() function can also shuffle along the first axis of a multidimensional array.

Example: Shuffle a 2D Array

arr_2d = np.array([[1, 2], [3, 4], [5, 6]])
shuffled_2d = np.random.permutation(arr_2d)
print(shuffled_2d)

Output:

[[5 6]
 [1 2]
 [3 4]]  # Example output

Note: Only the rows are shuffled; the internal structure of each row remains the same.

4. Difference Between permutation() and shuffle()

  • np.random.permutation(): Creates a new permuted array and leaves the original array unchanged.
  • np.random.shuffle(): Shuffles the array in place, modifying the original array.

Example: Using np.random.shuffle()

arr = np.array([1, 2, 3, 4, 5])
np.random.shuffle(arr)
print(arr)

Output:

[3 5 1 4 2]  # Example output

5. Reproducibility with Random Seed

To reproduce the same random permutation, use np.random.seed().

Example: Set a Seed for Reproducibility

np.random.seed(42)
shuffled = np.random.permutation(10)
print(shuffled)

Output:

[8 1 5 0 7 2 9 4 3 6]

Use Cases of Random Permutations

  • Shuffling Datasets:
    Randomize rows in a dataset for machine learning. 
data = np.array([[1, 'A'], [2, 'B'], [3, 'C']])
shuffled_data = np.random.permutation(data)
print(shuffled_data)
  • Games:
    Shuffle cards or tiles in games.
  • Simulations:
    Create randomized simulations in research.
  • Testing Algorithms:
    Test sorting algorithms with randomized inputs.

Summary

Random permutations in NumPy allow you to shuffle arrays and sequences easily, enabling randomized experiments, data preparation, and simulations. The numpy.random.permutation() function is versatile, offering a simple yet powerful way to handle randomization in your projects.

Explore more Python tutorials on The Coding College to take your coding skills to the next level!

Leave a Comment