A comparative study of K-Means, DBSCAN and OPTICS
Hari Krishna Kanagala, V. V. Jaya Rama Krishnaiah · 2016
In view of today's information available, recent progress in data mining research has lead to the development of various efficient methods for mining interesting patterns in large databases. It plays a vital role in knowledge discovery process by analyzing the huge data from various sources and summarizing it into useful information. It is helpful for analyzing the volumes of data in different domains like Marketing, Health, Science and Technology. Cluster analysis is widely used approach to notice the trends in the volumes of data. In this paper, we evaluated the performance of the different clustering approaches like as K-Means, DBSCAN, and OPTICS in terms of accuracy, outlier's formation, and cluster size prediction.