PSO-based Prediction Method for CNN-LSTM-Attention Networks

In the information age, the open nature of the Internet has made network security issues increasingly severe. Within the domain of network security, predicting the security situation is a fundamental aspect, which aims to identify potential security threats. To effectively quantify the current state of network security, we develop a situational assessment system that evaluates the impact of network attacks. Meanwhile, in order to improve the accuracy of situation prediction, a CNN-LSTM-Attention network prediction method based on particle swarm optimization algorithm is proposed. Simulation experiments demonstrate that the average absolute error of our proposed model is 0.0321 and its determination coefficient is 0.9660, indicating a notable improvement in prediction accuracy.