Hybrid prediction method for network traffic

Abstract

The invention discloses a hybrid prediction method for network traffic. The method is characterized in that a network traffic prediction method with an extreme learning machine (ELM) being compensated by autoregressive integrated moving average (ARIMA), namely, a network traffic prediction method with the ELM being compensated by a Farctal autoregressive integrated moving average model is provided through a self-similarity analysis of a network traffic sequence. The method comprises the following steps: firstly, predicting the network traffic sequence with the ELM; secondly, correcting an error sequence of network traffic prediction through an ARIMA model; and lastly, overlaying an ELM predicted value and an ARIMA model correction value to obtain a final predicted value. According to the method, prediction error data is fitted by the ARIMA model, and the predicted value of the ELM is overlaid with a residual error of ARIMA prediction to obtain the final predicted value. Residual error compensation is performed through the ARIMA model, thereby effectively increasing the prediction accuracy.

Description

technical field [0001] The invention belongs to the technical field of computer networks, and in particular relates to a method for predicting network traffic. Background technique [0002] Network traffic is an important parameter of current network management. When designing a network congestion control strategy when network resources are limited, accurate prediction of network traffic is essential for reducing network congestion, rationally allocating resources, improving network service quality, and discovering abnormal network behaviors. has a very important role. Research in recent years has found that network traffic shows certain changing rules even when the traffic changes suddenly, which makes it possible to analyze and predict network traffic sequences. [0003] At present, some research results regard network traffic as a linear model, respectively adopting autoregressive moving average (ARMA) model (see LanerM, SvobodaP, RuppM. ), differential autoregressive m...

AI Technical Summary

Problems solved by technology

However, with the increase of network complexity, the characteristics of network traffic have exceeded the Poisson or Markov distribution in the traditional sense. Therefore, there are theoretical deficiencies in using linear models for prediction, and it is difficult to guarantee the accuracy of prediction.

The prediction of nonlinear models mainly includes support vector machines (see LiaoWJ, BalzenZ. LSSVM network flow prediction based on the self-adaptive genetic gorithmoptimization [J]. Journal of Networks, 2013, 8(2): 507-512), artificial neural network (see WangJS, WangJK, ZhangMZ. Prediction of internet traffic based on Elmanneural network [C] / / Chinese Control and Decision Conference, 2009:1248-1252) and gray model (see Sun Hanlin, Jin Yuehui, Cui Yidong, etc. Gray model prediction of coarse-grained network traffic [J]. Journal of Beijing University of Posts and Telecommunications, 2010,33(1) :7-11) and so on, although the prediction accuracy of the nonlinear model has been improved to a certain extent compared with the linear model, but the neural network has the disadvantages that it is easy to fall into the local optimal value and the network structure is difficult to determine

Although the support vector machine requires a small number of samples, its key parameters are difficult to determine. The gray model is only suitable for cases where the data changes are not drastic, so the nonlinear model prediction of network traffic is difficult to guarantee the prediction accuracy

Images