Download PDFOpen PDF in browserA combined clustering algorithm based on ESynC algorithm and a merging judgement process of microclustersEasyChair Preprint 145544 pages•Date: September 4, 2019AbstractESynC algorithm is inspired by SynC algorithm and a linear version of Vicsek model. When facing complex data distributions, ESynC algorithm may regard a whole irregular cluster as some microclusters. In order to conquer this shortcoming, a Combined clustering algorithm based on ESynC algorithm and a merging judgement process of microclusters (CESynC) is presented. CESynC algorithm uses ESynC algorithm to detect clusters or microclusters and a merging judgement process to merge those conjoint microclusters. For some data sets that ESynC algorithm and SynC algorithm cannot detect correct clusters, CESynC algorithm can obtain natural clusters. From some experiments of some artificial data sets, we observe that parameter δ in CESynC algorithm has better valid interval than ESynC algorithm and SynC algorithm in some cases. From the experiments of nine artificial data sets, we observe that the valid interval of parameter σ is affected by parameters δ and MinPts. From the experiments of eight UCI data sets, we observe that CESynC algorithm gets better (or the same) clustering results than (or as) that of ESynC algorithm. From many experiments, we observe that the clustering results of CESynC algorithm and ESynC algorithm are often better than that of SynC algorithm. So we can say CESynC algorithm can often obtain better clustering quality than ESynC algorithm and SynC algorithm in some kinds of data sets. Further comparison experiments with some classical clustering algorithms demonstrate the clustering effect of CESynC algorithm. Keyphrases: ESynC algorithm, SynC algorithm, merging judgement, microcluster, synchronization clustering
