5G Traffic Sort-Term Prediction Based on OVMD Noise Reduction Decomposition and Hybrid CNN- BIGRU Optimized by Attention Mechanism

With the popularization and development of 5G networks, the performance requirements of mobile networks are getting higher and higher. Accurate 5G traffic prediction is of great help to network operation and maintenance management and daily network detection. Aiming at the nonlinear and non-stationary characteristics of 5G traffic data, this paper proposes a traffic prediction model based on OVMD noise reduction decomposition and hybrid depth CNN-BIGRU-ATTENTION model. First, the optimized VMD model decomposes the historical traffic data into stationary series to reduce data noise. Then, CNN-BIGRU-ATTENTION is constructed to predict the decomposed components one by one. CNN is used for feature extraction, BIGRU is used for time series fitting, and ATTENTION is used to focus on the influence of information. Finally, the final result is obtained by integrating components. To verify the validity of the model, this paper selects the Italian laboratory data as an example for analysis. The experimental results show that the model is superior to other comparison models and has the best flow prediction performance.