A Study of Clustering Approaches and Validation Measures for Big Data Mining | Original Article

ABSTRACT:

Present time natures of data are totally changed to big data with respect to volume, variety, and velocity. Clustering is one of the unsupervised approaches of data mining and data mining is one of the approaches for big data analysis and known as big data mining. Clustering is a very helpful technique under big data mining because it discovers distribution of patterns, hidden relations, self-noise and outlines management, class label predication and interesting correlations in large data sets and high dimensional data. Every traditional clustering algorithm works under specific criteria and these criteria define the cluster and validation measure validates this cluster as the requirements. From a theoretically, practically and the existing research perspective, this paper study seven clustering taxonomies such as partition, hierarchical, density, grid, model, fuzzy and graph based clustering taxonomy and their validation measures for the big data mining.