网络流量预测对网络安全、网络管理等具有重要的意义。针对网络流量的行为特征，提出了基于小波变换的PCNN网络流量预测算法。对预处理的网络流量进行小波分解，利用PCNN模型预测获得的近似系数和细节系数，通过小波逆变换对预测的小波系数进行重构，得到预测的网络流量。实验结果表明，与其他的三种网络流量预测算法相比，算法得到较小的残差，取得了较好的预测效果。%Network traffic prediction is very important for network security, network management and so on. According to network behavior characteristics of network traffic, an improved network prediction model is proposed based on wavelet transformation and PCNN. In this paper, a wavelet transformation is needed to the preprocessing network traffic in advance. Then PCNN is conducted to get the similarity coefficient and detail coefficient. The predicting network traffic is obtained by reconstructing the predicting wavelet coefficients with the inverse of wavelet transformation. Experimental results show that the method is superior to the other three methods with smaller residual and better predicting results.
展开▼
