Identity Recognition Scheme Based on User Access Behavior

Users generate a large amount of access data in the process of network interaction, and their access behavior contains rich personal information that can represent the user identities. How to effectively extract personality characteristics from these access behaviors, and establish a user access model, is the key to recognize the user identities. The existing recognition methods are mainly based on face, fingerprint, gait and other biological characteristics, with high recognition accuracy but high requirements for data acquisition. This paper was based on user history behavior data that is easily available, proposed the method of time window partition and feature extraction based on fuzzy representation, fusion of multiple features to obtain semantically rich web usage context for model training, and finally establishes a personal access model with multi-layer perception for identity recognition. Experimental results show that the proposed method is accurate in real data sets.