The algorithm used to determine the class labels was a stacked classifier using a Radial Basis Function Network (Netlab) and a Support Vector Machine (SVMTorch). First, the rbf network was trained on and subsequently labelled the training set, and the svm learned the training set with the precomputed attributes plus the output of the rbf network for the current point and the seven previous points. During testing, a sample is first labelled by the rbf network, and then the svm predicts the class label based on the attributes of the current point and the rbf labels of the current and seven previous points.