Effective feature selection with Particle Swarm Optimization based one-dimension searching

Forming an efficient feature space for classification problems is a grand challenge in pattern recognition. Many optimization algorithms are adopted to do feature selection, but these algorithms do searching in multi-dimensions space and always cannot get the optimal feature subset. In this paper, a feature selection method with Particle Swarm Optimization based one-dimension searching is proposed to improve the classification performance. Experimental results show that the proposed method can do feature selection more effectively than the compared method and get much higher classification accuracy.