Feature Selection using Multi-Objective Clustering based Gray Wolf Optimization for Big Data Analytics

Although numerous efforts have been made to develop feature selection framework which is efficient in Big Data technology, complexity of processing big data remains a significant barrier. As a result, the computational complexity and intricacy of big data may block the data mining process. The feature selection method means, a required pre-processing approach to minimize dataset dimensionality for great advanced features and classifier performance optimization. In order to increase performance, feature selection are regarded to constitute the core of big data technologies. In recent years, many academics have moved their focus to data science and analytics for application scenarios leveraging integrating tools of big data. People take quite some time to engage, when it comes to big data. As a consequence, in a decentralized system with a high workload, it is crucial in making feature selection dynamic and adaptable. Multi objective optimal strategies for feature selection are provided in this work. This research adds to the creation of a strategy for enhancing feature selection efficiency in large, complex data sets. In this paper, a multi-objective clustering-based gray-wolf optimization algorithm (MOCGWO) is proposed for classification problems. Five datasets were used to show the robustness of proposed algorithm. The result analysis was compared with other optimization methodology such as GWO and PSO. This shows efficacy of MOCGWO algorithm.