An Intelligentized Method Of Network Security Situation Assessment

With the constant evolution of the Internet hacker technology, and the rapidly popularized in Internet, the network security is constantly being challenged, the potential threats faced is more and more large, the traditional passive defense system, such as Intrusion Detection System (IDS), Firewall, Anti-virus Software and etc., has failed to meet people needs about network security. The appearance of network security situation awareness(NSSA), which meet people needs in some degree, which includes the Situation perception, Situation comprehend and Situation prediction, NSSA do systemic research about extracting and fusing from the vast amounts of the network data, assessing the quality of current network and forecasting the future situation. NSSA plays an important role in promoting the defensive performance of the network, and strength the management of network security.In this article, we do the further study in the frame, data refinement, and situation assessment and situation prediction. To begin with, we do some analysis and research from the overall framework of the network security situation awareness in this article, in a result, further improve situational awareness of network security performance. Moreover, because the network data are large and redundant, this paper introduces Manifold Learning and Kernel Match Pursuit to reduce the dimensions of the network data, and then clustering, in this way, reducing the amount of data, and so there are prepare foundation for evaluating the network security situation rapidly. What's more, in situation assessment according to the similarity of human antibody concentration and the network attacks, we make use of artificial immune concentration and the level of analysis to assess the network security situation, and through experiments to validate the effectiveness of this method. Finally, in situation forecast, this paper propose Support Vector Machine base on PSO, which method solve the subjectivity of parameters selection of SVM, the simulation results show that the optimized Support Vector Machine forecasting model can forecast the network security situation with high accuracy.