Tackling Ant Colony Optimization Meta-Heuristic as Search Method in Feature Subset Selection Based on Correlation or Consistency Measures

摘要

This paper introduces the use of an ant colony optimization(ACO) algorithm, called Ant System, as a search method in two wellknownfeature subset selection methods based on correlation or consistencymeasures such as CFS (Correlation-based Feature Selection) andCNS (Consistency-based Feature Selection). ACO guides the search usinga heuristic evaluator. Empirical results on twelve real-world classificationproblems are reported. Statistical tests have revealed that InfoGain is avery suitable heuristic for CFS or CNS feature subset selection methodswith ACO acting as search method. The use of InfoGain is shown to bethe significantly better heuristic over a range of classifiers. The resultsachieved by means of ACO-based feature subset selection with the suitableheuristic evaluator are better for most of the problems comparingwith those obtained with CFS or CNS combined with Best First search.