The Research Of Identity Recognition Based On Mobile Users' Mobile Internet Flow

As security issues of mobile terminals are increasingly significant, it takes more attentions in the field of identification and authentication research. The identification and authentication recognition techniques mainly include explicit user identification and implicit identification. explicit user identification includes authentication mechanisms based on PIN/pass-lock, draw-a-secret, face recognition and fingerprint recognition.It's obviously we need implicit and continuous identity authentication, for it doesn't affect users' experience. Several implicit identity sensing approaches have been proposed in the past that leverage the sensors on mobile devices such as accelerometer, GPS, and touchscreen. But it's influenced by the performance of the phone, and current studies rely on users to show specific information, also rely on the sensitivity of the sensor, therefore most studies only stay on the user classification levels.A model based on the identification of mobile phone users' network habits is proposed in this paper. It works through the feature extracted from the users' daily cellular network flow. I use the improved algorithm of Fisher and KS2 D to extract the user features in the subdivision of users' frequent flow by time and application.The end of this thesis use several users data from mobile operators to verify the accuracy of the algorithm, compared to other implicit verification of identity, the method has lower sampling costs. Algorithm testing has been completed now, the feature library needs more efforts to finish.