空天预警研究学报2025，Vol.39Issue(6)：415-418,435,5.DOI:10.3969/j.issn.2097-180X.2025.06.005
基于改进极光优化的CNN-BiLSTM网络流量预测方法
A network traffic prediction method based on IPLO-CNN-BiLSTM
摘要
Abstract
Network traffic prediction is a key technology for achieving rational allocation of network resourc-es and optimizing network performance.In order to improve the prediction accuracy of network traffic,this paper proposes a method based on improved polar lights optimization(IPLO)-convolutional neural network(CNN)-bidi-rectional long short-term memory neural network(BiLSTM).Firstly,by integrating the CNN spatial local feature extraction and the BiLSTM long-range dependency capture characteristics is constructed a CNN-BiLSTM net-work traffic prediction model.Then,the dynamic reverse learning strategy based on Beta distribution is intro-duced into the PLO algorithm to solve its problems of premature convergence and falling into local optima.Final-ly,the IPLO algorithm is introduced into the CNN-BiLSTM to optimize the hyperparameters and obtain the opti-mal hyperparameter combination for the CNN-BiLSTM.The experimental results show that,compared with other network traffic prediction methods,the IPLO-CNN-BiLSTM method achieves the best prediction results in the re-al network traffic set.关键词
网络流量预测/卷积神经网络/双向长短期记忆神经网络/改进极光优化算法Key words
traffic prediction/convolutional neural network(CNN)/bidirectional long short-term memory neural network(BiLSTM)/improved polar lights optimization(IPLO)分类
信息技术与安全科学引用本文复制引用
苏玉泽,范雄华,石卉..基于改进极光优化的CNN-BiLSTM网络流量预测方法[J].空天预警研究学报,2025,39(6):415-418,435,5.基金项目
国家自然科学基金项目(62101591) （62101591）
