A Novel RNN Model with Enhanced Behavior Semantic for Network User Profile

User profile is an essential task especially in zero trust network. It can be used to evaluate the security level of the system by separating the normal users from the attackers. In this paper, we propose a network user profile method based on semantic behavior modelling and Recurrent Neural Network (RNN). The user behavior in the network is enhanced with semantic detection and the semantic features include source, target, content and time. The semantic mapping is implemented with the predefined tags, and the overall behavior trajectory of the user in the network will be further classified to different types (normal users or attackers) with RNN algorithm. According to the experiment results, the proposed method shows a good performance in the user profile task. The overall accuracy is increased by 17% compared with the method without semantic information.