Download PDFOpen PDF in browserA shrinking synchronization clustering algorithm based on a linear weighted Vicsek modelEasyChair Preprint 143538 pages•Date: August 27, 2019AbstractClustering is an unsupervised learning method that tries to find some distributions and patterns in unlabeled data sets. Although clustering algorithms have been studied for many years, none of them is all purpose. This paper presents a Shrinking Synchronization Clustering (SSynC) algorithm by using a linear weighted Vicsek model. It is inspired by Synchronization Clustering (SynC) algorithm and Vicsek model. After some analysis and comparison, we find that SSynC algorithm based on the linear weighted Vicsek model has better synchronization effect than SynC algorithm based on an extensive Kuramoto model and has similar synchronization effect with Effective Synchronization Clustering (ESynC) algorithm based on another linear version of Vicsek model. In the simulations, several clustering algorithms (SynC, ESynC, KMeans, FCM, AP, DBSCAN, and Mean Shift) are used as comparative algorithms. By some simulated experiments of some artificial data sets, several real data sets, and three picture data sets, we observe that SSynC algorithm not only gets better local synchronization effect but also needs less iterative times and time cost than SynC algorithm. Moreover, SSynC algorithm needs less time cost than ESynC algorithm and almost get the same local synchronization effect and the same iterative times. Extensive comparison experiments with some class clustering algorithms demonstrate the effectiveness of our algorithm. At last, it gives some research expectations to popularize this algorithm. Keyphrases: A linear weighted Vicsek model, Clustering, Kuramoto model, Shrinking synchronization, SynC algorithm, near neighbor points
