Research On Intrusion Detection System Based On Pattern Recognition

Intrusion detection technology is a more in-depth multi-level network protection means, which is also the hot spots in the security technology research recently. In this paper, by using of pattern recognition technology, we have designed and implemented a new AdaBoost-based intrusion detection methods. And the choice of a weak classification algorithm - RBF neural network had been discussed in great detail, and we also successfully construct a reasonable structure of the neural network which is suitable for intrusion detection. Before intrusion detection, we first use the isometric mapping algorithm changing data from high-to low-dimensional by dimensional mapping, to improve detection of the recognition rate and efficiency. The experiments results based on comparison between AdaBoost method and the RBF neural network prove that AdaBoost method can greatly improve the recognition rate of weak classification algorithm.