Parameter determination and feature selection for back-propagation network by particle swarm optimization

The back-propagation network (BPN) is a popular tool with applications in a variety of fields. Nevertheless, different problems may require different parameter settings for a given network architecture. A dataset may contain many features, but not all features are beneficial for classification by the BPN. Therefore, a particle-swarm-optimization-based approach, denoted as PSOBPN, is proposed to obtain the suitable parameter settings for BPN and to select the beneficial subset of features which result in a better classification accuracy rate. A set of 23 problems with a range of examples and features drawn from the UCI (University of California, Irvine) machine learning repository is adopted to test the performance of the proposed algorithm. The results are compared with several well-known published algorithms. The comparative study shows that the proposed approach improves the classification accuracy rate in most test problems. Furthermore, when the feature selection is taken into consideration, the classification accuracy rates of most datasets are increased. The proposed algorithm should thus be useful to both practitioners and researchers.