Instrusion Detection System Of Neural Network

This article analyses the actual reality of network security and briefly intruduces a few of aggressive measures which threaten the safety of network. From these we can see, the passive ways such as firewall and anti-virus software to combat network attack can not meet the increasingly needs of network accessing .As a result, an intrusion detection system must be established to differentiate between the legal accessing and inbreaking attack. That is intrusion detection.From the third chapter of this article , we know what intrusion detection system is about and several main intrusion detection systems that is popularly used nowadays, as well as advantages and disadvantages. With the rapidly progression of network and the increasingly needs of network accessing, the present intrusion detection measures based on regulation can not meet the requirement of network intrusion detection .Consequently, a new intrusion detection system is greatly in the need to settle the present question of network security. In the fourth chapter of this article , an intrusion detection measure based on statistics is clarified detailedly. After a serial of precise analysis, a conclusion can be made that this measure can detect the attacking actions of network and abusiveness of user's right. However, with the inherent calculating complexity of statistic measures , it is difficult to detect network accessing instantaneously. We bring forward neural network combining statistic measures to solve the instantaneous detection of network accessing. In the fifth chapter, the first step is to create training data aggregate of neural network with the use of statistic measures, then use this data aggregate to train BP neural network, after that , extract subcharacter including auditing records and user accessing, then , use Self-Organizing Mapping neural network to map the character of user to a vacuum of planar character, therefore , complete the extraction of accessing character, the last step is to use this network to complete the instantaneous detection of user accessing.In the end,I design an intrusion detection system based on neural network and sums up my study . In brief, the intrusion detection technology based on the neural network is an important direction in the environment of high-speed network presently .