randomly select element(s) from numpy array

Numpy – Select Random Elements From Array

In this tutorial, we will look at how to select one or more random elements from a Numpy array with the help of some examples.

How to select random elements from a Numpy array?

randomly select element(s) from numpy array

You can use the numpy.random.choice() function to select random elements from a given 1-D Numpy array. The following is the syntax –

# randomly select value(s) from numpy array
numpy.random.choice(a, size=None, replace=True, p=None)

It returns the selected random value or array of values (if selecting more than one random value).

The numpy.random.choice() function takes the following parameters –

  • a – The 1-D array to sample from. You can also pass an integer in which case the sampling will be done from the result of numpy.arange(a)
  • sizeOptional argument. The number of samples to draw. You can also pass a tuple (m, n, k) as the output shape, in which case m*n*k samples are drawn. Its default value is None, in which case a single value is sampled.
  • replace – Optional argument. Whether to sample with replacement. Its default value is True meaning samples are drawn with replacement (the same value can be sampled multiple times) by default.
  • pOptional argument. The array of probabilities associated with each value in a. If not specified, then sampling is done by assuming a uniform distribution over each value in a (they have the same probability of being sampled).

Examples

Let’s now look at some examples of using the above syntax to sample random elements from a Numpy array.

First, we will create a Numpy array that we will be using throughout this tutorial.

import numpy as np

# create a numpy array
ar = np.array([1, 2, 3, 4, 5])
# display the array 
ar

Output:

array([1, 2, 3, 4, 5])

Here, we used the numpy.array() function to create a 1-D array of some integers.

📚 Data Science Programs By Skill Level

Introductory

Intermediate ⭐⭐⭐

Advanced ⭐⭐⭐⭐⭐

🔎 Find Data Science Programs 👨‍💻 111,889 already enrolled

Disclaimer: Data Science Parichay is reader supported. When you purchase a course through a link on this site, we may earn a small commission at no additional cost to you. Earned commissions help support this website and its team of writers.

Example 1 – Select one random element from a Numpy array

If you only want to get a random value from a 1-D Numpy array, pass the array as an argument to the numpy.random.choice() function. We don’t need to specify the size argument because the function by default samples a single value.

Let’s sample a single value from the array created above.

# select a random value from ar
np.random.choice(ar)

Output:

4

We get a random value from the array ar.

Example 2 – Select multiple random values from a Numpy array

You can also sample multiple random values using the numpy.random.choice() function. Use the size parameter to specify the number of values you want to randomly sample from the array.

Let’s randomly sample 4 values from the above array.

# select 4 random values from ar
np.random.choice(ar, size=4)

Output:

array([4, 2, 1, 2])

We get a 1-D Numpy array with the randomly sampled elements.

You can see that the value 2 occurs twice in the resulting array. This is because the numpy.random.choice() function samples values with replacement by default. This means that a value can be sampled multiple times.

Example 3 – Select multiple random values (without replacement) from a Numpy array

If you want to sample without replacement (that is, a value cannot be chosen again if it’s already been sampled), pass replace=False to the function.

Let’s now randomly sample 4 values without replacement from the array ar.

# select 4 random values from ar without replacement
np.random.choice(ar, size=4, replace=False)

Output:

array([1, 2, 4, 3])

We get a 1-D array of the resulting random values (which have been sampled without replacement).

Example 4 – Select random values with a custom probability distribution

In the above examples, each value in the array ar had an equal probability of being sampled. You can pass a custom probability distribution (a 1-D array with associated probabilities for each value in ar) to the p parameter.

Currently, each value in the array ar has a 20% (0.2) probability of being selected. Let’s double the probability of the value 3, and proportionately reduce the probabilities of the other values. The probability distribution now looks like [0.15, 0.15, 0.40, 0.15, 0.15].

Let’s now use this probability distribution to randomly select 4 values (with replacement) from the array ar.

# select 4 random values from ar with custom probability distribution
np.random.choice(ar, size=4, p=[0.15, 0.15, 0.4, 0.15, 0.15])

Output:

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

You can see that the resulting array has multiple occurrences of 3. This is because the probability of selecting a 3 is much higher (40%) compared to the other values (15% each) in the array.

Summary

In this tutorial, we looked at how to randomly select values from a Numpy array. The following are the key takeaways from this tutorial.

  • Use the numpy.random.choice() function to randomly select values from a Numpy array.
  • Use the size parameter to specify the number of values to sample.
  • The numpy.random.choice() function samples the values with replacement by default. To sample without replacement, pass replace=False.
  • By default, each value in the array has an equal probability of being sampled (uniform distribution). To specify a custom probability distribution for the values, use the p parameter.

You might also be interested in –


Subscribe to our newsletter for more informative guides and tutorials.
We do not spam and you can opt out any time.


Author

  • Piyush Raj

    Piyush is a data professional passionate about using data to understand things better and make informed decisions. He has experience working as a Data Scientist in the consulting domain and holds an engineering degree from IIT Roorkee. His hobbies include watching cricket, reading, and working on side projects.

Scroll to Top