IoT individual privacy features analysis based on convolutional neural network

In terms of the evaluation problems about individual privacy defense situation of cyberspace, an evaluation method for individual privacy security defense situation based on convolutional neural network is proposed in this thesis; firstly, the state of situation factor at different times in the individual privacy defense system is subject to fuzzy and probabilistic processing based on Bayesian algorithm, to build up situation awareness and situation estimation model and then input the initial condition probability, state transition probability and observation data into the model; secondly, the convolutional neural network is introduced to recognize and evaluate the Bayesian algorithm model mentioned above, improving the accuracy of global restriction protection algorithm for individual privacy features; finally, simulation experiment is carried out to verify the performance advantages of modeling of the algorithm mentioned above. (C) 2018 Elsevier B.V. All rights reserved.