A multiobjective ACO algorithm for rough feature selection

Rough set theory has been widely applied to feature selection. In this paper, a multi-objective ant colony optimization algorithm is proposed for rough feature selection. This algorithm evaluates the constructed solutions on the basis of Pareto dominance. Moreover, it only uses the non-dominated solutions to add pheromone so as to reinforce the exploitation and adopts crowding comparison operator to maintain the diversity of the constructed solutions. In addition, it avoids premature convergence by imposing limits on pheromone values. Numerical experiments are carried out on gene expression datasets. Compared with a modified non-dominated sorting genetic algorithm, our algorithm can provide competitive solutions efficiently for rough feature selection.