Support Vector Machine Explainer
A support vector machine, or SVM, is a supervised method that can classify examples by finding a boundary with a wide margin between classes. The training examples closest to that boundary are the support vectors; they have a direct role in determining where the boundary sits.
When classes are not perfectly separable, a soft-margin SVM balances a wider margin against penalties for violations. The C parameter controls that tradeoff. Feature scaling matters because distances and margins can otherwise be dominated by variables with large numeric ranges.
A kernel lets an SVM represent nonlinear boundaries by computing similarity as if examples had been mapped into a richer feature space. Common kernels include linear, polynomial, and radial-basis functions. Kernel and parameter choices strongly influence both accuracy and computation.
SVMs can work well in high-dimensional settings and with moderate datasets, but training can become expensive as data grow. Raw decision scores are not automatically calibrated probabilities. Clear evaluation, scaling, parameter tuning, and comparison with simpler baselines are essential.
A support vector machine finds a decision boundary with a wide margin between classes, with selected examples defining that boundary.
Key terms include support vector, margin, separating hyperplane, soft margin, C parameter, kernel, radial-basis function, feature scaling and decision score.
The model searches for a boundary that separates classes while leaving as much space as possible to the nearest influential examples.
Allowing some violations makes the method usable when data overlap or contain noise instead of demanding a perfectly clean split.
A kernel measures similarity in a way that permits curved boundaries without explicitly constructing every transformed feature.
Kernel settings can overfit or underfit, so they should be selected with cross-validation and evaluated on an untouched test set.
Explore more "Explainers"
Discover additional explainers across politics, science, business, technology, and other fields. Each explainer breaks down a complex idea into clear, everyday language—helping you better understand how major concepts, systems, and debates shape the world around us.