Density clustering based on border-expanding
Dongming Chen, Yunhui Yan, Dongqi Wang · 2014
DBSCAN is a clustering algorithm based on density. It can divide regions which have a high density for clusters, shield the noise effectively and discover clusters of arbitrary shape and any size from dataset. However, DBSCAN algorithm needs to traverse dataset to find core objects, so it results in large amount of I/O cost when processing large-scale datasets. A fast algorithm (BEDBSCAN) is developed which expands the cluster by employing border objects as seeds. Experimental results show that BEDBSCAN performs obvious efficiency improvement than DBSCAN algorithm especially when processing large datasets.