An hybrid metaheuristic approach for efficient feature selection

Several new challenges as well as specialized difficulties are getting accumulated for big data that are against both scholarly research groups as well as and business IT sending. The rich big data sources are set up on information streams as well as the dimensionality scourge. It is difficult to precisely assess these big data for decision making systems. In the recent times, several domains are handling big datasets in which there is large number of additional features. The main aim of feature selection techniques is to eliminate noisy, redundant, or unrelated features that cause poor classification performance. This research implements the Feature selection employing Information Gain, Bacterial Foraging Optimization (BFO) as well as Hybrid BFO to compute on big data. Outcomes on various data sets reveal that the suggested Naïve Bayes, KNN method performs better when compared to the method analyzed in the literature.