Feature selection optimization through enhanced Artificial Bee Colony algorithm

Feature Selection extracts the more informative and distinctive features from any dataset to improve the classification accuracy. Evolutionary and swarm intelligent algorithms play a vital role in optimizing the process of FS. Artificial Bee Colony algorithm is a popular swarm intelligent, metaheuristic search algorithm and it has been widely used in solving numerical optimization problems. In our previous work, we had adapted the ABC algorithm as such and had proposed a new algorithm for FS (ABC-FS) and it had provided optimal subset of features. In this paper, we have made enhancements to the ABC algorithm to adapt and maintain the history of the previously abandoned and the global best solutions for FS optimization (EABC-FS). The enhanced ABC algorithm has been tested on 10 standard datasets and experimental results show the promising behavior of the proposed algorithm.