基于ResNet-LSTM组合模型的网络流量预测研究

Research on Network Traffic Prediction Basedon ResNet-LSTM Combined Model

对网络流量的准确预测,不仅是网络安全稳定运行的保障,还是运营商合理调度网络资源的重要参考.为了提高网络流量预测精度,提出一种基于残差网络与长短时记忆网络相结合的流量预测方法.首先,使用残差卷积层提取原始数据特征,并将提取的特征向量输入LSTM各节点,然后,LSTM细胞单元通过循环连接进行长序列预测,最后,通过输出层输出预测结果.利用淮南汽车站采集到的网络流量数据进行实验仿真,并与卷积网络、残差网络和长短时记忆网络预测方法对比,实验结果表明,ResNet-LSTM模型预测精度更高.

Accurate prediction of network traffic is not only a guarantee for network security and stable operation,but also an important reference for operators to reasonably schedule network resources.In order to improve the accuracy of network traffic prediction,a traffic prediction method based on the combination of residual network and long short-term memory network is proposed.The residual convolutional layer is used to extract the original data features,and the extracted feature vectors are input to each LSTM node,and the LSTM cell units are predicted by cyclic connections,and finally the prediction results are output through the output layer.The network traffic data collected by Huainan Bus Station is used for experimental simulation,and compared with the prediction methods of convolutional network,residual network and long-term short-term memory network,the experimental results show that the ResNet-LSTM model has higher prediction accuracy.