微型电脑应用2025，Vol.41Issue(5)：26-28,32,4.
基于PSO-Attention-BiLSTM的网络安全态势预测模型构建研究
Research on the Construction of Network Security Situation Prediction Model Based on PSO-Attention-BiLSTM
摘要
Abstract
To improve the accuracy of network security situation prediction,a prediction method based on PSO-Attention-Bi-LSTM network is proposed.By using an improved particle swarm optimization(PSO)algorithm with adaptive inertia weight factor and acceleration factor,the optimal learning rate,random seeds and the number of neurons in the first and second hidden layers of the birdirectional long short-term memory(BiLSTM)network are solved.The attention mechanism is introduced to distinguish between key and non-key features,and the BiLSTM network is improved.The PSO-Attention-BiLSTM network is used to predict the network security situation.The results show that the mean absolute percentage error(MAPE)and symmet-ric mean absolute percentage error(SMAPE)predicted by the improved BiLSTM network are 1.34 and 0.15,respectively,with a coefficient of determination(R2)of 0.99.Compared to commonly used models such as random forest,autoregressive in-tegrated moving average(ARIMA)model and grey prediction,the predicted values of the preposed method are closer to the true values,and the proposed method has significant advantages in various performance indicators.It can be concluded that the preposed method can more accurately predict network security situations,and has certain practical reference value.关键词
网络安全/态势预测/双向长短期记忆网络/粒子群优化算法/注意力机制Key words
network security/situation prediction/BiLSTM network/particle swarm optimization algorithm/attention mecha-nism分类
信息技术与安全科学引用本文复制引用
胡慧霞..基于PSO-Attention-BiLSTM的网络安全态势预测模型构建研究[J].微型电脑应用,2025,41(5):26-28,32,4.基金项目
河南省教育科学规划一般课题(2020YB0107) （2020YB0107）
