基于自组织映射神经网络的局部自回归方法在网络流量预测中的应用

发布时间：2018-06-26 04:50

本文选题：网络流量 + 预测　；参考：《信息与控制》2016年01期

【摘要】：针对网络流量预测,提出一类基于自组织映射(self-organizing map,SOM)神经网络的局部自回归(auto-regressive,AR)方法.根据SOM的联想记忆在时域的推广,在矢量量化临时联想记忆(vector-quantized temporal association memory,VQTAM)建模技术的基础上,给出具有多个局部线性AR模型的AR-SOM方法,基于前K个获胜神经元用权值代替输入向量建立单一时变局部AR模型的K-SOM方法,以及在完成数据向量聚类的同时,更新多个局部AR模型系数的LLM(local linear map)-SOM方法.相对于全局模型,基于SOM神经网络的局部AR方法能够灵活给出有效的监督神经结构,降低了计算复杂度.将本文方法应用于不同的网络流量预测实例中,并与现有方法相比,实验结果表明所提出的方法能有效地改善预测精度,且性能更好.
[Abstract]:For network traffic prediction, a class of auto-autoregressive (AR) method based on self-organizing mapSOM neural network is proposed. According to the extension of SOM's associative memory in time domain, based on the modeling technique of vector-quantized temporal association memory (VQTAM), an AR-SOM method with multiple local linear AR models is presented. The K-SOM method based on the weights of the first K winning neurons instead of the input vectors to establish a single time-varying local AR model, and the LLM (local linear map) -SOM method to update the coefficients of multiple local AR models at the same time as the data vector clustering is completed. Compared with the global model, the local AR method based on SOM neural network can provide the effective supervised neural structure flexibly and reduce the computational complexity. The proposed method is applied to different network traffic prediction examples. Compared with the existing methods, the experimental results show that the proposed method can effectively improve the prediction accuracy and the performance is better.
【作者单位】：兰州交通大学自动化与电气工程学院;
【基金】：国家自然科学基金资助项目(51467008)
【分类号】：TP393.06;TP183

【相似文献】