A variant of DBSCAN algorithm to find embedded and nested adjacent clusters

In this paper we present an efficient approach for clustering analysis to detect embedded and nested adjacent clusters using concept of density based notion of clusters and neighborhood difference. Basically our proposed algorithm is improved version basic DBSCAN algorithm, proposed to address the clustering problem with the use global density parameters in basic DBSCAN algorithm and problem of detecting nested adjacent clusters in EnDBSCAN algorithm. Our experimental results that suggested that proposed algorithm is more effective in detecting embedded and nested adjacent clusters compared both DBSCAN and EnDBSCAN without adding any additional computational complexity. Also we have preset method to evaluate the global density parameters using sorted k-distance plot and first order derivative.