An improved intrusion detection framework based on Artificial Neural Networks

Faced with high dimensional and large amount of data, network intrusion detection is always the focus of current research in the network security field. With the advantages of nonlinear, distributed storage and easily computing, Artificial Neural Networks (ANNs) are widely used in machine learning and pattern recognition fields. In this paper, we adopt a feature selection algorithm based on Fisher to select feature subsets, and three typical neural network algorithms for classification in order to improve the results of the intrusion detection. Experiments adopt KDD´99 as the data set, and use the accuracy, false positive rate and false negative rate, to evaluate the feasibility and effectiveness of the three neural networks. And as a result, the experiments show that the algorithms have acceptable performance in intrusion detection.