Abstract:

With the problems of the low detection rate and the insufficient sensitivity to the new intrusion，the current intrusion detection systems affect the functions of the entre system. Based on the very deep research, this paper proposes a new neural network ensembles method for intrusion detection. The method is used to train the individual networks on the basis of data reduction. Neural network techniques are used to combine the different classification results. Theory and experiment show that the model is effective.

Key words: neural network ensembles;attribute selection;intrusion detection