Abstract:

Aiming at the dynamic and real-time characteristics of network security situation prediction,a network security situation prediction model based on hidden Markov model（HMM）was designed.According to the HMM,the situation in t+1 from t time can be calculated.Meanwhile,the maximum entropy algorithm was combinated with HMM to improve the accuracy rate.The analysis and simulation experiments show that the method has the characteristics of autonomous learning and active defense,and it also balance the speed and accuracy.

Key words: network security situation prediction, hidden Markov model, regularization, network security