Research And Implementation Of Network Security Situation Awareness Model Based On Machine Learning

Network Security Situation Awareness(NSSA)can build a comprehensive and unified situation assessment system for complex cyberspaces,and can predict the dynamic evolution of future security risks,it provides security strategies for the entire network.Therefor,the NSSA has become one of the new hot spots in the field of network information security.Based on the knowledge of machine learning,this thesis conducts in-depth research on network security and situation assessment and prediction mothods.The main contributions in this thesis are as follows:(1)In order to improve the accuracy of situation assessment,a network security situation assessment model based on an improved artificial fish swarm algorithm optimized multi-class twin support vector machine is proposed in this thesis.The artificial fish swarm algorithm is applied to the parameter optimization process of the twin support vector machine.When dealing with the multiclassification problem,the clustering algorithm is combined with the binary tree multi-classification to reduce the cumulative error in the multi-classification process.The model can accurately evaluate the current network security situation.Through experiment,the model is compared with the existed network security situation assessment models,which verifies the superiority of the model in the accuracy of the situation assessment.(2)In order to improve the accuracy of network security situational awareness prediction,a network security situation prediction model based on radial basis function(RBF)neural network optimized by hybrid hierarchical genetic algorithm is proposed in this thesis.The analysis gives a nonlinear mapping between the network security situation value which has the feature of time serial and nonlinear.In the method,the topology and parameters of RBF neural network are optimized by hybrid hierarchy genetic algorithm and the global search ability of genetic algorithm is improved by simulated annealing algorithm.The model can accurately predict the future situation.Through experiment,the proposed model is compared with the existed network security situation prediction models,which verifies the superiority of the model in the accuracy of situation prediction.