SVM Intro

less than a minute

Geometric Intuition

We have two classes of points (e.g. Cats vs. Dogs ) that can be separated by a straight line.
Many such lines exist !

SVM asks: “Which line is the safest?”

Highway ️ Analogy

Think of the decision boundary as the center-line of a highway.

SVM tries to make this highway as wide as possible without hitting any ‘buildings’ (data points) on either side.

Support Vectors

The points that lie exactly on the edges of the highway are the Support Vectors.

Goal 🎯: Maximize the width of the ‘street’ (the margin) to ensure the model generalizes well to unseen data.

End of Section