Based On SOM Intrusion Detection Algorithm Of Research And Improvement

With the development of network applications, the using of network information transmission have become increasing frequently, so the network information security has become the focus of our attention. The safety of network information is the foundation of Internet and information technology development, so how to make the network more illegal conduct defense initiative and effectively restrain today, has become the important problem in network security research. A number of existing security protection technology and product are useing passive safety protection methods for network information protection, limitations, and intrusion detection system for network information transmission to surveillance, real-time transmission can find suspicious alarms or take active response measures, so use intrusion detection technology to strengthen the network information protection is an important means of present application.The detection method based on existing rules are used by the traditional intrusion detection system, this has caused that some system only get better detection results to specific or already known breaches, but can not adapt to the complex network environment,and can not effectively detect the unknown intrusion activities affect system as a whole, detection efficiency. Therefore introducing neural network technology, using neural network technology on data processing ability, enhance the intrusion detection accuracy, make the intrusion detection system realizing better detection effect,.This paper is based on SOM neural network of network intrusion detection algorithm for data, and by using the test method of learning vector quantization to improve, to increase SOM neural network on the intrusion detection efficiency of data, so as to improve the SOM neural network based on the intrusion detection system performance. Finally the application of MATLAB simulation software and security audit data sets on the learning vector quantization of improved SOM network intrusion detection simulation test and analysis of the results. Through the test data prove that the improved network intrusion detection method in improving the intrusion detection performance is feasible and effective.