Find in Library

Network intrusion detection is one of the most important issues of network security and is a research interest of many researchers. In this paper, we present a model based on the combination of recurrent neural networks and rule sets for the network intrusion detection problem. The main idea of the model is to combine the strengths of each classification model. The rule set is capable of detecting known attacks, while the recurrent neural network has the advantage of detecting new attacks. A comparison of the detection efficiency of our model with the previous detection models on the same data set, KDD CUP 99, shows that the proposed model is effective for detecting network intrusions at rates higher than 99%.