Categorical › Classification (supervised)

Support vector machine (SVM)

A support vector machine separates classes with a maximum-margin boundary, using kernels for nonlinear separation, reported with 10-fold cross-validated accuracy.

What is Support vector machine (SVM)?

SVMs find the separating boundary that maximizes the margin to the nearest training points (the support vectors). The cost parameter C trades margin width against training errors; kernels (RBF by default) let the boundary bend.

Predictors are standardized internally. The 10-fold CV accuracy is the honest number; training accuracy is optimistic, especially with flexible kernels.

When should I use Support vector machine (SVM)?

Medium-sized classification problems with smooth nonlinear boundaries.
High-dimensional predictors relative to n.

What data does it need?

Categorical outcome + numeric predictors + kernel + cost.

What does it report?

10-fold CV accuracy, training accuracy, support-vector count, confusion matrix.

What does it assume?

Independent observations.
Results depend on C / kernel choice — compare CV accuracy across settings.

How do I interpret the result?

A large train-vs-CV accuracy gap = overfitting; lower C or use a simpler kernel.

References

Cortes & Vapnik (1995). Support-vector networks. Machine Learning 20.