How to Convert Index Arrays to One-Hot Encoded Arrays in NumPy?

One-Hot Encoding Index Arrays in NumPy

In NumPy, converting a 1D array of indices to a one-hot encoded 2D array is a common task. For example, given the array a with indices [1, 0, 3], we want to encode it as:

b = [[0,1,0,0], [1,0,0,0], [0,0,0,1]]

To achieve this, there are two key steps:

1. Create a zeroed array: Create a 2D array b with enough columns (i.e., a.max() 1) to accommodate the one-hot encoded values. The array should be initialized with zeros.
2. Set appropriate values to 1: For each row i in b, set the a[i]th column to 1. This indicates that the original index a[i] is present at position i in the one-hot encoded array.

Here's a code example to illustrate:

<code class="python">import numpy as np

a = np.array([1, 0, 3])
b = np.zeros((a.size, a.max() + 1))
b[np.arange(a.size), a] = 1

print(b)</code>

Output:

[[0. 1. 0. 0.]
 [1. 0. 0. 0.]
 [0. 0. 0. 1.]]

This method effectively converts the array of indices into a one-hot encoded array, where each row represents a one-hot encoded value of the corresponding index in a.

The above is the detailed content of How to Convert Index Arrays to One-Hot Encoded Arrays in NumPy?. For more information, please follow other related articles on the PHP Chinese website!