A survey of neural network ensembles

A neural network ensemble combines a finite number of neural networks or other types of predictors, which are trained simultaneously for a common classification task. Compared with a single neural network, the ensemble is able to efficiently improve the generalization ability of the classifier. The objective of this paper is to introduce existing research work on the neural network ensembles, including effective analysis, general implement steps of ensembles, and traditional technologies for training component neural networks, and also description the applications of it.