Optimizing weights by genetic algorithm for neural network ensemble

Combining the outputs of several neural networks into an aggregate output often gives improved accuracy over any individual output. The set of networks is known as an ensemble. Neural network ensembles are effective techniques to improve the generalization of a neural network system. This paper presents an ensemble method for regression that has advantages over simple weighted or weighted average combining techniques. After the training of component neural networks, genetic algorithm is used to optimize the combining weights of component networks. Compared with ordinary weighted methods, the method proposed in this paper achieved high predicting accuracy on five test datasets.