Research On Network Security Situation Element Extraction And Prediction Technique

With the widespread use of Internet and the booming growth of various applications,the accompanying network security risks are gradually increasing.The existing security defense technology only guarantees one aspect of network security and lacks coordination mechanism between each other.Under such circumstances,the network security situation awareness,as an active defense technology,has rapidly become one of the top topics of current research.Situational awareness extracts the situation factors from a large number of multi-source heterogeneous security state data in order to do real-time evaluation and prediction of network security.In this thesis,adaptive enhancement algorithm,improved particle swarm algorithm and artificial neural network are combined to study the acquisition and prediction of network security situation.The concrete are as follows:Firstly,in order to extract the information of network security situation accurately and effectively,a hierarchical frame feature acquisition method based on enhanced probabilistic neural network is proposed.According to different functions of Agent node,the hierarchical feature acquisition framework is divided into different levels.The principal component analysis(PCA)is used to reduce the training sample attributes and the special attribute encoding fusion.The result can be used to optimize the structure of the probabilistic neural network(PNN)so as to reduce the system complexity.Then,the improved PNN is used as the base classifier.Combined with the adaptive enhancement algorithm,the final strong classifier is formed through repeated iteration,weight replacement and weighted fusion.The experimental results show that the proposed model achieve higher accuracy and better generalization ability than other methods.Secondly,in order to accurately grasp the development trend of network security situation,a situation forecasting method based on IPSO-WNN is established according to the dynamic and non-linear features of network security information.First,the prediction model is established by wavelet neural network(WNN)with strong nonlinear ability and generalization performance,and then the improved particle swarm optimization algorithm is used to optimize the model parameters.Due to the fact that the particle swarm optimization(PSO)is characterized by the problem of slow convergence speed and easily getting into the local optimum,the optimization method in this thesis in the iterative process combines the mutation factor in the genetic algorithm,the affinity propagation(AP)and Gaussian weighted global extremes,which improves the global search ability of algorithm and the performance of the model.The simulation results indicate that the proposed method can improve the convergence speed of the algorithm and the accuracy of the situation prediction.