Comparison of SVM and SVDD on a simple classification problem, where the sampling is balanced (a) and imbalanced (b). The shading indicates the true classification, and the points are the training data.