Research On Network Security Situation Prediction Method Based On Improved Grey

With the rapid development of network and communication technology, the scale of Internet is getting more and more larger and widely used in many fields. It also makes people work and life convenient. At the same time, the occurring frequency of network security events is greatly increasing. Furthermore, these events appear organized and strong pointed. Traditional safety protection equipments are largely used. But they are independent of each other and don’t protect the security of network from the whole point of view. In order to provide reference opinions for network administrator making countermeasures in time, network security situation awareness assess the whole current network security state from macro perspective and forecast the future trend of security state. Concrete research of this paper includes the following two aspects.Firstly, on the bases of acquiring network security factors, These factors were analyzed by association and matched by rule in order to fuse in classification into situation assessment index. Hierarchical network security situation assessment index system covering all security factors is established. These network weights are determined by the entropy weight method. Then, the security situation was evaluated. A improved Particle Swarm Optimization(PSO) algorithm through embedding Genetic algorithm into PSO is put forwarded in order to improve the precocious problem of PSO algorithm. Then, considering the BP neural network’s problems of low efficiency and not ideal parameters optimization results in training process, the improved PSO was used to train the BP neural network. On MATLAB simulation experiment platform, the improved PSO was applied to train the neural network compared with the PSO algorithm and BP algorithm. The experiment proved the improved PSO had higher efficiency and better index optimization results and corresponding neural network acquired better forecasting results.A network security situation forecasting method was proposed based on the corrected grey by BP neural for the shortcomings of traditional forecasting methods. Considering the advantage of grey only needing less date to establish a model and easy to be operated, grey neural network model was used to forecast network security situation to gain initial situation forecasting value. The strong nonlinearity handle ability and nice pattern identification ability of BP neural were used to correct the initial situation forecasting value. In order to acquire really date, a network attack defense environment based on vmware virtual honeynet was created. On the MATLAB simulation environment, the model of this article is applied to forecast network security situation compared with other models. Experiment results verify the method have higher accuracy.