How to Convert an Array of Indices to a One-Hot Encoded Array in NumPy?

Converting an Array of Indices to a One-Hot Encoded Array in NumPy

Often, it becomes necessary to transform a 1D array of indices into a 2D array where each row represents a one-hot encoding of the corresponding index in the original array.

Example:

Let's have a 1D array of indices 'a':

<code class="python">a = np.array([1, 0, 3])</code>

We aim to create a 2D array 'b' where each row is a one-hot encoding of the corresponding index in 'a':

<code class="python">b = np.array([[0, 1, 0, 0], [1, 0, 0, 0], [0, 0, 0, 1]])</code>

Solution:

To achieve this transformation, we can utilize the following steps:

1. Zeroed Array Creation:
   Create a zeroed array 'b' with enough columns to accommodate the maximum index value in 'a', plus one ('a.max() 1').

<code class="python">b = np.zeros((a.size, a.max() + 1))</code>

2. One-Hot Encoding:
   For each row 'i' in the array, set the 'a[i]'th column to 1. This step transforms each index in 'a' into a one-hot encoded row in 'b'.

<code class="python">b[np.arange(a.size), a] = 1</code>

Output:

Executing this code produces the desired one-hot encoded array 'b':

<code class="python">[[ 0.  1.  0.  0.]
 [ 1.  0.  0.  0.]
 [ 0.  0.  0.  1.]]</code>

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