Computer network security evaluation method based on improved attack graph

Abstract

ABSTRACT The traditional security detection and defense methods are relatively backward, and there are many problems, such as high false alarm and missed alarm rate, and detection lag, which have been difficult to meet the requirements of computer security defense. In order to strengthen the active defense of computer network security and improve the classification and prediction performance of computer network security events, a computer network security risk assessment model based on mrmr Ig feature selection model and attack graph model is proposed. The mrmr Ig feature selection model is used to classify the characteristics of security events, and the hidden Markov chain model and attack graph model are used to predict the attack intention. The experimental results show that the classification accuracy rate of the model is 99.79%, the false alarm rate and the false alarm rate are 0.11% and 0.18%, respectively. The model can accurately predict attacks in real time, and can provide reference for timely taking targeted computer network security reinforcement and active defense measures.