Enhancing decision-based neural networks through local competition

In this paper, the decision-based neural network (DBNN) learning algorithm is modified to stimulate local competition. Performance is assessed in ten UCI databases, resulting in improved results at the expense of a relatively low increase of the computational burden.