Feature Selection Based on Asynchronous Discrete Particle Swarm Optimal Search Algorithm
Wen-Ting Hsieh, Shi‐Jinn Horng · 2012
The feature subset selection reduces the cost of collecting redundant features. It is the main goal of feature subset selection that generating a feature subset which can preserve the most useful information of the original features. The feature selection methods often need expensive cost to find the optimal feature subset. The asynchronous discrete particle swarm optimal search algorithm is proposed to implemented and applied in the feature selection. The experimental results show that the proposed algorithm outperforms the others with respect to effective and efficient. The contributions of this study are: to survey methodology for feature selection, to apply the ADPSO-based algorithm on feature selection, and to construct an evaluated function for feature selection.