SVM

An SVM (Support Vector Machine) is a classifier that finds the decision boundary separating two classes with the maximum margin — the widest gap between the boundary and the nearest points (the support vectors). The kernel trick lets it draw nonlinear boundaries by implicitly mapping data to a higher-dimensional space.

Worked example: to separate two classes a straight line cannot split, an RBF-kernel SVM finds a curved boundary; only the support vectors (points near the boundary) determine it, so the rest of the data does not matter. Gotcha: SVMs shine on small-to-medium, clean, high-dimensional datasets (text classification was a classic use), but they scale poorly to very large data (training is roughly quadratic) and need careful kernel and regularization (C) tuning — which is why gradient boosting and neural nets displaced them on big data.