Network Security Situation Prediction Based on CPSO-WaveNet

A cross-particle swarm optimization approach is suggested that improves the WaveNet network security situation prediction model (CPSO-WaveNet) in order to increase the model's prediction accuracy and speed. Using WaveNet's advantageous nature of easy extraction of time series data, the features among network security situation data are learned, and accurate prediction is performed. Meanwhile, the particle swarm optimization (PSO) algorithm is improved by introducing a vertical crossover operation to enhance population diversity. The model's hyperparameters are optimized using crossover particle swarm optimization (CPSO). In the experiments, on the one hand, the effectiveness of the crossover particle swarm algorithm is verified by two standard test functions. On the other hand, the simulation experiments and the established prediction model are compared with the traditional prediction method, and the proposed prediction method can achieve a fit of 0.9908. Comparing the convergence speed to the comparison algorithm, it is noticeably faster.