• Valarmathy.N
• Dr.Krishnaveni Sakkarapani

The primary goal of data mining is to organize data into meaningful clusters using several clustering algorithms. The DBSCAN algorithm, which may be used for a variety of applications, is the most effective of the numerous clustering techniques. This technique is a high-quality density-based method with various advantages, including the ability to identify arbitrarily shaped clusters, and the ability to identify outliers and ignore them before clustering. Epsilon (Eps) and a minimal number of points (MinPts) are the two key input factors that have a significant impact on clustering performance. To overcome this problem, the K-distance graph approach is used to automatically choose Eps and MinPts, and spatial access methods are used to speed up neighbourhood calculation for each data point. The suggested new technique makes use of the spatial access method and the k-distance graph method to improve scalability and accelerate the execution process. The results of the experiments clearly show that combining the K-Distance tree method with the DBSCAN algorithm is efficient in terms of all metrics and can cluster both high and low dimensional data effectively.

DBSCAN, Outlier detection, Speed Optimization, Nearest Neighbour Search

"IMPROVISATION OF DBSCAN CLUSTERING ALGORITHM USING VARIOUS METRICS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.9, Issue 6, page no.6-16, June-2022, Available :http://www.jetir.org/papers/JETIRFN06002.pdf

"IMPROVISATION OF DBSCAN CLUSTERING ALGORITHM USING VARIOUS METRICS", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.9, Issue 6, page no. pp6-16, June-2022, Available at : http://www.jetir.org/papers/JETIRFN06002.pdf