Abstract
The big data comprises a relatively developing area of study which is due to numerous facts gathered daily and the wishes to be helpful information for use in our day by day life. One of the most crucial pre-processing of data is the feature selection. This paper proposes a hybrid technique that combines two algorithms namely Gray Wolf Optimization (GWO) and Particle Swarm Optimization (PSO), the manner that lets in the critical functions to be recognized and lets the insignificant ones and the complexity to be erased. This enables the obligations of the gadget learning classification while making use of training to the classifier with the data set. A hybrid approach is primarily based on metaheuristics swarm intelligence algorithms, which simulate the gray wolf's management and hunting manner in nature and PSO which people are moving impacted by their local best positions and by the global best position. This hybridization is to acquire the balance between exploitation and exploration. We used seventeen datasets from UCI machine gaining knowledge of repository within the experiments and comparisons results to assess the effectiveness and quality of the GWOPSO.