Iterative Label Spreading
Amanda Susan Barnard, Parker, Amanda · 2019
Based on a general definition of a cluster and the quality of a clustering result, this code presents a new method for evaluating existing clustering algorithms, or undertaking clustering, capable of predicting the number and type of clusters and outliers present in a data set, regardless of the complexity of the distribution of points. This algorithm, referred to as iterative label spreading (ILS), can recognize the characteristics expected of a successful clustering result before any clustering algorithm has been applied, providing a type of hyper-parameter optimization for clustering. In this notebook the algorithm, is assessed using large benchmark two-dimensional synthetic data sets, with tutorial examples.