Research on network security situation awareness algorithm based on deep learning

Existing network monitoring methods are often only capable of single monitoring and cannot effectively respond to the current complex and diverse network attacks. Network security situation awareness has become a key research direction. Based on CNN and exponentially weighted D-S evidence theory, a multi-source fusion network security situation awareness algorithm is designed to comprehensively monitor the network. The decision engine uses CNN as the attack recognition engine, making full use of the feature extraction and feature fusion functions of the convolution kernel to improve the ability to accurately identify attacks; multi-source fusion uses exponentially weighted D-S evidence theory to effectively fuse the output results of each decision engine, further improving the attack recognition performance; situation assessment uses weight coefficient theory to effectively quantify the threat level, and uses hierarchical analysis methods to accurately obtain the security situation of the entire network. Experimental results show that the proposed situation awareness algorithm improves the attack recognition accuracy to 92.76%, which is better than most published research results. The improvement in accuracy helps to more accurately calculate the security situation of the network.