Feature Selection with Particle Swarm for Improved Classification on High-Dimensional Datasets

The feature selection (FS) problem, being complex and computationally resourceful, poses a bigger challenge when handling data in the high-dimensionality realm. Swarm intelligence (SI) is a new way to approach the most complex problems. SI has been steadily growing in its popularity and it has been demonstrated that it has the ability to handle a wide range of optimization problems, e.g., feature selection. It is proven by the fact of its application in many areas. Swarm intelligence algorithms in recent times have garnered popularity to a great extent which is mainly because of their wide search space and simple methodology. One of the methods that we will consider is Particle Swarm Optimization (PSO) for feature selection which may enhance the performance of k-Nearest Neighbors (KNNs) classifier for datasets with high dimensions. Also, our study involved an in-depth comparison and analysis between the PSO feature selection technique and other optimization techniques.