Label Propagation Algorithm

Label propagation is a semi-supervised algorithm in machine learning that assigns labels to previously unlabeled data points, starting from a small subset of points that already carry labels and propagating those labels to the rest over the course of the algorithm. Within complex networks it is also used as a community-detection algorithm, exploiting the tendency of real networks to have community structure, with an advantage in running time and in requiring no prior knowledge of network structure, though its disadvantage is that it produces no single unique solution.