Fig. 1From: Regularized Simple Graph Convolution (SGC) for improved interpretability of large datasetsThe application of the proposed methodology to the circular data. Data points are produced about the origin at a fixed radius so that 100 points are equally spaced. The data points have class labels allocated so that there are 50 in each of 2 classes. This is shown in the plot with a line which determines the separationBack to article page