Research on service model of prediction in cognitive networks

How to provide great service to users is one of the most important jobs to cognitive networks. Cognitive networks can perceive the external environment; intelligently and automatically change its behaviour to adapt the environment. This feature is more suitable to provide security for users with Quality of Service. This paper proposes a hybrid traffic prediction model, which trains BPNN with Ant Colony Algorithm based on the analysis of the present models. Furthermore, the model includes three stages, and the model predicts the network traffic with the hybrid model. The proposed model can avoid the problem of slow convergence speed and an easy trap in local optimum when coming up with a fluctuated network flow. Thus, the traffic prediction with high-precision in cognitive networks is achieved. © Dandan et al.