Support Vector Machine (SVM)

Harvard EPS-210 | Interactive tutorial — Explore maximum margin classification

Classification Space & Margin

Class -1
Class +1
Support Vector

Decision Function f(x)

f(x) = w·x + b
Classify: sign(f(x))

Hinge Loss Visualization

Training Progress

Add Data Points

Margin Width

--
2 / ||w||
0 Support Vectors

Regularization (C)

C Parameter 1.00

Small C → wider margin, more errors
Large C → narrower margin, fewer errors

Kernel Function

Model Statistics

Accuracy
--
Loss
--
Class -1
0
Class +1
0

Quick Examples

How it works: SVM finds the hyperplane that maximizes the margin between classes. Support vectors are the critical points that define this margin. The kernel trick enables nonlinear boundaries.