Solving Feature Subset Selection Problem by a Hybrid Metaheuristic.
Miguel García-Torres, Félix César García López, Belén Melián-Batista, José Andrés Moreno Pérez, José Marcos Moreno-Vega · 2004
The aim of this paper is to develop a hybrid metaheuristic based on Variable Neighbourhood Search and Tabu Search for solving the Feature Subset Selection Problem in classification. Given a set of instances characterized by several features, the classification problem consists of assigning a class to each instance. Feature Subset Selection Problem selects a relevant subset of features from the initial set in order to classify future instances. The proposed hybrid metaheuristic is compared with a Genetic Algorithm proposed in the literature. Although he hybrid metaheuristic and the genetic algorithm had a similar performance according to the accuracy percentages, the hybrid metaheuristic provided a higher reduction in the set of features.