ML CodingmediumNumpy~15m

One-Hot Encoding

Problem

You are preprocessing categorical labels for a neural network classifier. Neural networks cannot directly use integer class labels like 0, 1, 2 — they need a binary vector representation where only one position is "hot" (set to 1).

Given a 1D array of integer class labels and the total number of classes n_classes, return a 2D one-hot encoded matrix of shape (len(labels), n_classes). Row i should have a 1 at column labels[i] and 0 everywhere else.

Always use n_classes for the number of columns — the input labels may not cover all classes.

Examples

Example 1

Input: ([0, 1, 2], 3)

Output: [[1, 0, 0], [0, 1, 0], [0, 0, 1]]

Each label maps to its column. Result is a 3x3 identity matrix when labels are [0,1,2].

Example 2

Input: ([0, 1], 5)

Output: [[1, 0, 0, 0, 0], [0, 1, 0, 0, 0]]

n_classes=5 but only 2 labels. Trailing columns are zeros — always trust n_classes not max(labels)+1.

Example 3

Input: ([2, 2, 2], 4)

Output: [[0, 0, 1, 0], [0, 0, 1, 0], [0, 0, 1, 0]]

All same class — identical rows. Each row is independent, no state carries between rows.

Constraints

›1 <= len(labels) <= 1000
›0 <= labels[i] < n_classes
›2 <= n_classes <= 100

Reference solution

Reference solution available after you attempt the question.

Ready to solve it?

Start a session on Mockbit #72. Write your code, run it against hidden tests, and get graded with specific critique on each axis.