A Reduct Computation Approach Based on Ant Colony Optimization
Tamara Qablan, Qasem A. Al‐Radaideh, Sawsan Abu Shuqeir · 2012
Rough set theory provides an important concept for feature reduction called reduct. The cost of reduct set computation is highly influenced by the attribute set size of the dataset where the problem of finding reducts has been proven as an NP-hard problem. Ant Colony Optimization (ACO) as a meta-heuristic technique has been successfully applied to several combinatorial problems. This paper proposes an approach for reduct computation based on ACO methodology. The proposed approach has three main features: (1) the updated pheromone trials are directed to the nodes that visited by the ants rather than the visited edges connecting these nodes; (2) the pheromone trial values are limited between max and min trial limits; and (3) the heuristic value is evaluated dynamically during the ant search. To verify the proposed approach, several experiments are carried out on nine standard UCI datasets. . The results of experiments have showed that the proposed approach can produce a short reduct with less number of iterations in comparison to other ACO based feature reduction approaches.