| Nowadays,the internet industry continues to develop,and various software and websites with a lot of users have been born.The demand for security has also increased while the rapid growth has occurred,and network security has become a hot issue for all people.The intrusion traffic detection problem and malicious user screening problem are discussed in this paper.In the latest data analysis research,knowledge graph has new data analysis features that can analyze the potential connections between data to perform knowledge inference.Therefore,we propose a systematic solution for securing the network based on the study of knowledge graph.The main work of this paper:(1)A knowledge extraction algorithm for traffic data is proposed.Most of the previous algorithms use neural networks,pattern matching and other methods for knowledge extraction,which cannot be applied to the processing of traffic data.Therefore,based on the characteristics of traffic data,we process traffic data on the basis of clustering algorithm and completes the knowledge extraction of traffic data in this paper.(2)A knowledge inference method for intrusion detection on access traffic is proposed.Based on the existing knowledge graph model,the knowledge processing of many-to-one relational data is optimized,the triadic information of traffic data is successfully modeled,and the modeled knowledge graph database is used for knowledge inference to judge whether the access traffic is intrusion traffic.(3)A modeling method for simulating user relationships in complex social networks is proposed.Firstly,the dimension selection strategy in the knowledge graph model is optimized,after which the intrinsic attribute characteristics of users and the relationship network among users are modeled using the characteristics of the knowledge graph to establish a knowledge graph database of user relationships,and then the malicious users in the user group are distinguished according to the knowledge graph database. |
PDF Full Text Request