Research on Intrusion Detection System Based on Integrated Transfer Learning

At present, Internet security has been paid more and more attention, but the websites of many institutions or organizations are still frequently subjected to abnormal network attacks. These attacks may cause the network to be unblocked and normal users to access the web page slower; heavy Otherwise, the website may be paralyzed so that it may be down for a long time. Since network attack behaviors often appear when a user sends a request to the server, analyzing and modeling the acquired link behavior characteristics and distinguishing between normal and abnormal network requests can be done as soon as possible Avoid abnormal network requests. In addition, since the network attack behavior is not static, it is particularly important to pre-warn the new network attack behavior in advance. This paper proposes a network attack behavior prevention strategy based on integrated transfer learning technology. The data is used as an auxiliary sample to establish an intrusion detection model under the new environment, which can detect and prevent network insecurity factors as soon as possible.