IPSO-LSTM: A New Internet Security Situation Prediction Model

Abstract:
Situation prediction is an important part of situation awareness. Because Internet devices will generate a large amount of heterogeneous multi-source data, most of these data are scattered and cannot be shared. It is difficult for network security personnel to make appropriate decisions about the network based on these data. Adjustment and decision-making are often passive in the game with hackers. In order to solve the above-mentioned problems, a situation prediction technology is proposed to predict the network situation for a period of time in the future through the analysis and understanding of the past network security situation factors. Network security data has the characteristics of time series, and long and short-term memory networks can usually process time series data more effectively. The improved PSO algorithm is used to automatically iteratively search for the optimal parameters to reduce human subjectivity. The improved particle swarm algorithm and the long and short-term memory network are effectively combined through encoding to improve the accuracy of the prediction method.