Neural networks ensemble approach for detecting attacks in computer networks

Network activity has become an essential part of daily life of almost any modern person or company. At the same time the number of network threats and attacks of various types in private and corporate networks is constantly increasing. Therefore, the development of effective methods of intrusion detection is an urgent problem at the present day. In this paper we propose a new approach to intrusion detection in computer networks based on the use of neural networks ensembles. This approach can be implemented in distributed intrusion detection systems (IDS) which better meets the challenges of the present time, in contrast to the traditional use of neural networks in host-based IDS. In the paper the basic steps of the neural networks ensembles designing are described and some of the methods to complete these steps are expounded. Peculiarities of using neural networks ensembles to solve classification problems are discussed. Then the basic scheme of neural networks ensemble approach to intrusion detection systems is proposed. Conditions and results of the experimental investigation of the proposed approach on a number of classification problems are presented, including the problem of classifying probe attacks. Possible development of the proposed approach and areas for future research are discussed in the end.