Research On Network Security Situational Awareness Based On LSTM-DT Model

In today's society,the striking expansion of Internet technology such as big data,cloud computing,and artificial intelligence have facilitated people's lives and promoted social production,but at the same time,they have also caused many security issues.As the fourth largest space in addition to sea,land and air,cyberspace's security is so important.Therefore,network security situational awareness technology has flourished,which can help us understand the actual operation of the network,and predict its future trend based on the historical state.In order to better understand the attacker's intention and describe the network state,this paper proposes a network security situational awareness model.The contents and innovations are as follows:·The concept of "attack probability" is proposed in this paper to predict the probability of various attacks through long-term and short-term memory neural network,which changes the limitation that previous scholars can only identify the types of attacks.· This paper proposes the concept of "probability of attack",which changes the limitation of previous scholars that situation awareness can only identify attack types,and makes the final situation value more accurate.· This paper puts forward the concept of "attack impact degree",uses analytic hierarchy process and decision tree algorithm to identify the attack types,and evaluates the attack impact,so that the identified attack types can be expressed more objectively.·The model is applied to UNSW?NB15 and KDD?CUP99 data sets to carry out experiments.The results show that the model has good stability.When describing the general network environment,the prediction accuracy can reach 95%,while when describing the complex network environment,the prediction accuracy can still be maintained at more than 80%.Combined with the occurrence probability and influence degree of early attack,the final network situation value is obtained,and the corresponding curve is drawn to analyze the network operation state.Experiments show that this method has some progress compared with previous research results.