基于小波变换和极限学习机的网络流量预测模型

为了提高网络流量预测准确性,结合网络流量的变化特点,针对当前网络流量预测模型存在的局限性,设计了基于小波变换和极限学习机的网络流量预测模型。首先分析了当前国内外网络流量预测研究现状,找到引起网络流量预测准确性差的原因;然后采用小波变换对原始网络流量时间序列进行去噪,得到无噪声的网络流量时间序列;最后采用极限学习机对网络流量时间序列进行建模,得到相应的预测结果。与当前经典的网络流量预测模型在相同环境下进行对照测试,测试结果分析表明,小波变换和极限学习机的网络流量预测精度达到了95%以上,网络流量预测误差得到了有效的控制,而且提升了网络流量预测效率,预测结果要远优于当前经典的网络流量预测模型。

Network Traffic Prediction Model Based on Wavelet Transform and Limit Learning Machine

In order to improve the accuracy of network traffic prediction, considering the changing characteristics of network traffic and the limitations of current network traffic prediction models, a network traffic prediction model based on wavelet transform and extreme learning machine is designed. Firstly, the current research status of network traffic prediction at home and abroad is analyzed, and the reasons for the poor accuracy of network traffic prediction are found. Then, wavelet transform is adopted. The original network traffic time series is denoised by transformation, and the network traffic time series without noise is obtained. Finally, the network traffic time series is modeled by the limit learning machine, and the corresponding prediction results are obtained. The test results are compared with the current classical network traffic prediction model at the same environment. The analysis of the test results shows that the network of the wavelet transform and the limit learning machine is the same. The accuracy of network traffic forecasting is over 95%, the error of network traffic forecasting is effectively controlled, and the efficiency of network traffic forecasting is improved. The forecasting result is much better than the current classical network traffic forecasting model.