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

With the progress and development of information technology,the network environment has gradually become huge and complex,and the types of attacks in the network have become more and more diversified.Network security has also developed from the original passive defense to the active prevention.Network security situation awareness technology is born in such an environment.Network security situation awareness technology can evaluate the current security situation of the network and predict the future development trend of the network,which has great research significance and practical value.This paper focuses on the situation assessment technology and situation prediction technology,aiming at the problem of strong subjective determination of weight in situation assessment method and low prediction accuracy in situation prediction model.In addition,the network security situation awareness model in this paper is designed,the situation assessment index system in this paper is formulated,and a set of network security situation awareness system is designed and implemented.The main work of this paper is as follows:1.The existing network security situation awareness model and index system are systematically studied and summarized Combined with the research of this paper,the network security situation awareness model of this paper is established,and the network security situation indicator system of this paper is established.2.Aiming at the problem that the subjectivity of weight determination is too strong in the existing network security situation assessment methods,the evidence reasoning assessment method based on combined weight is proposed.The weights of the current evidential reasoning method are combined with subjective and objective weights.Among them,calculate subjective weights through fuzzy analytic hierarchy process,and dynamically calculate objective weights based on evidence data.3.Aiming at the problems of low prediction accuracy and slow convergence speed in the existing network security situation prediction model,a network security situation prediction model based on bidirectional LSTM is proposed,and the super parameters of the model are determined by Bayesian optimization method,and the model is optimized by Xavier initialization weight method.Through the experiment proves the practicability and advantage of this method.4.Design and implement a network security situation awareness system.The architecture design,module division,detailed design and implementation are carried out.