Dealing Big Data using Improved Fuzzy C Means Based Improved Redundant Particle Swarm Optimization with Map Reduction

There is a lot of interest in big data analysis from both the academic and commercial worlds. A number of algorithms have been implemented to improve the analysis process. Improved Fuzzy C-means clustering has been used to improve the MapReduce model in this study. 'In this work, Redundant Particle Swarm Optimization with Multi-Objective Optimization (MOO-MR-RPSO) based Improved Fuzzy C Means (IFCM) clustering mechanism is used along with MapReduce model. Each data element is mapped together and forms the data heterogeneity attributes. Due to unbalanced data in the large datasets, it is necessity to extract the features of the data; it is performed by using the Principle Component Analysis (PCA) based optimal feature extraction and feature selection respectively. Finally, MOO-MR-RPSO based optimization mechanism is developed for selecting the both appropriate clusters with optimal centroid(s) in the IFCM by using PCA features based objective function and the method named as the MOO-IFCM-RPSO with map reduce. The simulation results shows that the proposed MapReduce approach gives Maximum clustering accuracy compared to the state of art approaches.