In the context of SVM$,$ what is the hinge loss function used for?

To calculate probabilities

To determine the depth of the decision tree

To measure the error for data points being on the correct side of the margin

To compute the distance between data points