基于神经网络的网络流量预测算法研究

网络流量预测有助于网络服务质量的提升和网络资源的合理分配,对优化网络管理与运营、保障用户体验质量至关重要。因特网业务的急剧增加和基础网络的快速发展导致网络流量变得更加复杂多样,传统网络流量预测模型难以保证较高的预测精度,而神经网络作为人工智能的重要分支,在预测复杂网络流量时具有显著优势。简述反向传播神经网络、径向基神经网络和长短期记忆神经网络的模型原理,通过分析这些神经网络预测不同时间尺度的网络流量结果,可总结其预测性能与优缺点,为基于神经网络的故障预测和故障定位的学术研究和实际应用提供技术支撑。

Research on Network Traffic Prediction Algorithms Using Neural Networks

Network traffic prediction is beneficial to the enhancement of network service quality and reasonable allocation of network resources,which plays an important role in optimizing network management and operation,and ensuring users’quality of experience(QoE).The sharp increase of Internet services and the rapid development of infrastructure networks lead to more complex and diverse network traffic,which makes it difficult for traditional network traffic prediction models to guarantee the higher prediction accuracy.As an important branch of artificial intelligence,neural networks have significant advantages in predicting complex network traffic.This paper briefly introduces the model principles of back propagation neural network(BPNN),radial basis function neural Network(RBFNN),and long short-term memory neural network(LSTMNN).It then summarizes their prediction performance,advantages and disadvantages by analyzing the network traffic prediction outputs that result from using them to predict network traffic at different time scales.The effort made in this paper is expected to provide technical support for academic researches and practical applications related to fault prediction and fault location based on neural networks.