Classification with K-Nearest Neighbours

Definition: Classification predicts a category (not a number). K-Nearest Neighbours (KNN) is a simple, intuitive classifier: to label a new point, it looks at the K closest known points and takes a majority vote.

Example — apple or orange?

Features are [weight, texture] where texture is 0 = smooth, 1 = bumpy:

from sklearn.neighbors import KNeighborsClassifier

# features: [weight, texture]   labels: the fruit
X = [[150,0],[170,0],[140,0],[130,1],[120,1],[110,1]]
y = ["apple","apple","apple","orange","orange","orange"]

model = KNeighborsClassifier(n_neighbors=3)
model.fit(X, y)

print("A 160g smooth fruit is:", model.predict([[160, 0]])[0])
print("A 115g bumpy fruit is: ", model.predict([[115, 1]])[0])

How KNN decides

For a new fruit, KNN finds the 3 most similar known fruits and picks the most common label among them. No complex maths — just "what does it most resemble?"

💡 Try it: change the test fruit values and re-run to see the prediction flip between apple and orange.