A Network Security Situation Prediction Method Based On Attention-CNN-BiGRU
Chuhang Hu, Guikai Liu, Ming Li · 2022 IEEE 25th International Conference on Computer Supported Cooperative Work in Design (CSCWD) · 2022
To improve the accuracy and efficiency of network security situation prediction, a network security situation prediction method based on Attention-CNN-BiGRU is proposed. It combined Convolutional Neural Network(CNN) and bidirectional Gated Recurrent Unit(BiGRU), The structure of the proposed model is more accurate for extraction and understanding of situation time series; besides, to improve the performance of proposed model, the Attention mechanism is utilized to optimize the model; particle swarm optimization(PSO) is applied to optimize varieties of hyperparameters of the model. The optimal combination of hyperparameters selected by PSO is used for training of the model. Finally, the method proposed in this paper was verified in experiment and compared with other models. The results shows that the method can acomplish the task of situation prediction with better capability.