DEBC-GM: Denclue Based Gaussian Mixture Approach for Big Data Clustering

In current digitization world, data are growing with high density rapid rate. Therefore, it is necessary to manage the complexity of data in an efficient way with less effort. In order to handle the complex data, a concept called clustering has been introduced to help the user to recognize the natural grouping of a dataset. The objective of the data clustering is to find the groups of identical substances in a dataset while preserving them separately from the noisy points. Different discrete clustering algorithms have been proposed to manage the consequences of the large datasets. To rectify the noisy instances of large datasets, in this paper, a clustering methodology named DENCLUE-GM is introduced. This methodology is framed on the connectivity and density functions. Preliminary results display that the clustering veracity rate, clustering quality, and the global search efficiency of improved algorithm is greater than classical clustering algorithms. On the other hand, DENCLUE-GM methodology is compared with DBSCAN approach in the aspects of accuracy, better memory utilization, and other quality based measures