ECR-DBSCAN: An improved DBSCAN based on computational geometry
Kinsuk Giri, Tuhin Kr. Biswas, Pritisha Sarkar · Machine Learning with Applications · 2021
A new density based clustering algorithm E C R − D B S C A N based on D B S C A N , has been presented in this paper. Computational geometry is applied to develop the modified D B S C A N algorithm. It is well known that the quality of density based clustering depends on its input parameters. However, it is not easy to determine proper values of input parameters for D B S C A N . This paper presents three significant modifications or extensions to D B S C A N related with (i) selection of hyper parameter e p s i l o n ( e p s ) using the radii of empty or voronoi circles (ii) selection of parameter minPoints ( m p ) for the same epsilon and (iii) redistribution of noise points to suitable clusters using the concept of centroid hinged clustering. E C R − D B S C A N is implemented with PYTHON accompanied by extensive experiment on benchmark data sets. Our experimental results establish the novelty and validity of the proposed clustering method over standard techniques.