Research On Network Security Situation Assessment Method With Deep Weighted Feature Learning

With the continuous development of the network in the information age,network security incidents caused by malicious attacks or destruction are more and more common.Network and information systems are facing the threat of many network attacks.Network security situation assessment network(NSSA)can identify,understand and assess various activities to grasp the security status of the entire network and support reasonable security response decisions.However,the available network security situation assessment methods have difficulties in the aspects of difficulty in feature extraction and low accuracy in attack detection.To tackle this problem,this thesis makes a detailed analysis of network security situation assessment methods and deep learning technology and further discusses how to apply deep learning to network security situation assessment.On this basis,this thesis proposed a network security situation assessment method with deep weighted feature learning.The proposed network security situation assessment method with deep weighted feature learning includes three parts: situation extraction,situation analysis,and situation assessment.The situation extraction is designed to process complex and diverse network traffic data into a format that meets the input requirements of the deep learning model.The situation analysis is designed to detect and analyze the processed data.To improve the threat detection effect,a PFEN-ABi GRU model composed of a parallel feature extraction network(PFEN)and an attention-based bi-directional gate recurrent(ABi GRU)network is designed.The model extracts the key information of different network threats differentially through the PFEN module and fuses the extracted features with the original information,and then inputs the fused features into the ABi GRU module to complete the weighting of the key features in the network traffic to achieve the purpose of fast and accurate detection of network threats.The situation assessment is designed to analyze and quantify the detection results obtained in the situation analysis process,and calculate the network security situation value to evaluate the network security situation.In the experimental part,the proposed method is verified by multiple experiments.The effectiveness of the PFEN-ABi GRU in the situation assessment module is verified by experiments,and the performance improvement effect of the PFEN-ABi GRU compared with other threat detection models in terms of precision,recall,and F value is proved.Through comparative experiments with typical network security situation assessment methods,the reliability and feasibility of the proposed method are verified.