How does SVM relate to distance-based learning models?

SVM is not distance-based; it is margin-based, unlike k-NN or clustering methods.

Distance-based models rely on proximity metrics for prediction. SVM instead constructs a global separating hyperplane that maximizes margin, making it less sensitive to local distance distortions.