Hybrid visual computing models to discover the clusters assessment of high dimensional big data

Clusters assessment is a major identified problem in big data clustering. Top big data partitioning techniques, such as, spherical k-means, Mini-batch-k-means are widely used in many large data applications. However, they need prior information about the clusters assessment to discover the quality of clusters over the big data. Existing visual models, namely, clustering with improved visual assessment of tendency, and sample viewpoints cosine-based similarity VAT (SVPCS-VAT), efficiently perform the clusters assessment of big data. For the high-dimensional big data, the SVPCS-VAT is enhanced with the subspace learning techniques, principal component analysis (PCA), linear discriminant analysis (LDA), locality preserving projection (LPP), Neighborhood preserving embedding (NPE). These are used to develop hybrid visual computing models, including PCA-based SVPCS-VAT, LDA-based SVPCS-VAT, and LPP-based SVPCS-VAT, NPE-based SVPCS-VAT to overcome the curse of dimensionality problem. Experimental is conducted on benchmarked datasets to demonstrate and compare the efficiency with the state-of-the-art big data clustering methods.