K-Nearest Neighbors Classifier

Classify a point by majority vote among its **K nearest training examples** — the simplest non-parametric classifier.

Algorithm

1. Compute Euclidean distance from query to all training points

2. Select the K nearest neighbors

3. Return the label with the most votes

Example

train = [[0,0],[1,0],[0,1],[1,1]]
labels = [0, 0, 1, 1]
knn(train, labels, query=[0.4,0.4], k=3)  →  0