Explanation

Option B is correct because support vector machines (SVMs) aim to find the optimal hyperplane that maximizes the margin between classes. The support vectors are the data points closest to the decision boundary, and the algorithm seeks to maximize the distance between these support vectors and the decision boundary.

Option A is incorrect because adding training data away from the support vectors typically does not affect the decision boundary, as the boundary is determined primarily by the support vectors themselves.

Option C is incorrect because in soft margin classification, the algorithm balances margin width and classification errors using a regularization parameter (C). It doesn't minimize total error penalty regardless of margin width; rather, it finds a trade-off between maximizing margin and minimizing classification errors.