Applying Ant Colony Optimization in configuring stacking ensemble
Yijun Chen, Man–Leung Wong · 2012
A stacking ensemble is a collective decision making system employing some strategy to combine the predictions of learned classifiers to generate its prediction on new instances. The early research has proved that a stacking ensemble is usually more accurate than any individual component classifiers both empirically and theoretically. Though many ensemble methods are proposed, it is still not an easy task to find a suitable ensemble configuration for a specific dataset. In some early works, the ensemble is selected manually according to the experience of the specialists. Metaheuristic methods can be alternative solutions to find configurations. Ant Colony Optimization (ACO) is one popular approach among the metaheuristics. In this paper, we propose a new ensemble construction method which applies ACO in the Stacking ensemble construction process to generate domain-specific configurations. Different kinds of local information are applied in facilitating the learning process. A number of experiments are performed to compare the proposed approach with some well-known ensemble methods on 18 benchmark datasets. The experiment results show that the new approach can generate better stacking ensembles.