ConDPC: Data Connectivity-Based Density Peak Clustering

As a relatively novel density-based clustering algorithm, Density peak (DPC) has been widely studied in recent years. DPC sorts all points descending order of local density and finds neighbors for each point turn to assign the appropriate clusters. The algorithm is simple effective but some limitations applicable scenarios. If difference between clusters large or data distribution nested structure, effect this poor. This study incorporates idea connectivity into original proposes an improved ConDPC. ConDPC modifies strategy obtaining center assigning improves accuracy algorithm. In study, comparison experiments were conducted on synthetic sets real-world sets. compared algorithms include DPC, DBSCAN, K-means two over DPC. results prove effectiveness