Figure 4From: Path classification by stochastic linear recurrent neural networksThe resulting hyperplanes obtained by ERM and SVM by introducing mislabelled training data (the dotted lines represent the hyperplanes obtained in the first experiment)Back to article page