One-Hot Encoding of Nominal Values

Problem

Implement one-hot encoding for nominal (categorical) values. Given a 1D array of labels, return a 2D binary matrix where each row corresponds to a sample and each column corresponds to a unique class.

Solution

import numpy as np

def one_hot_encode(labels):
    unique = sorted(set(labels))
    label_to_idx = {label: i for i, label in enumerate(unique)}
    n_samples = len(labels)
    n_classes = len(unique)
    encoded = np.zeros((n_samples, n_classes))
    for i, label in enumerate(labels):
        encoded[i, label_to_idx[label]] = 1
    return encoded

Explanation

1. Find all unique labels and sort them for deterministic column ordering.
2. Map each label to a column index.
3. Create a zero matrix of shape (n_samples, n_classes).
4. For each sample, set the entry at the corresponding class column to 1.

Complexity

• Time: O(n) where n is the number of samples
• Space: O(n * k) where k is the number of unique classes