Study On Network Traffic Analysis And Prediction Model

The Study of the characteristics of network traffic and the passive measurement of network traffic are basic methods and ways to deeply understand the essence of network,and to comprehend the operation of network.In this paper, prediction model based on neural network algorithm, introducing season cyclical chaos algorithm and decomposition model, build a new network traffic prediction model. According to the actual network of measured network traffic data, and simulation experiments, the results show that the new model and algorithm prediction error and low, and generality. In view of the network traffic prediction, this paper mainly do the following job:(1) Aimed at the short comings of static feed forward network and Elman network in network traffic prediction, new modified Elman neural network is proposed, and a leaning algorithm based on seasonal periodicity proposed, With a large amount of network traffic data from the actual network , on basic of which, the network traffic was predicted. Simulation experimental results show this model has better effect of prediction, Compared with traditional linear model, BP neural network model and Elman neural network model, it has higher precision and better adaptability. The result of predicted shows the model is feasible and reasonable, and it shows that the model is feasible and effective.(2) With a large amount of network traffic data collected from the actual network, a new modified Elman neural network is proposed. The chaos searching was introduced into the model training, the data redundancy was reduced and the problem of local optimum was effectively solved using ergodicity of the Tent map. Experimental results show that the new model and the strategy can improve the network's training speed and the forecast accuracy of network traffic.(3) Considering the characteristics of network traffic macro behavior, using mathematical tool will network flow timing is decomposed into structure relatively simple son composition, to describe and predict the network traffic behavior. Nonlinear rule Meanwhile based on neural network theory, this paper puts forward behavior prediction model proposed, including decomposition model and neural network model, fully considered flow behavior of periodic, trend and randomicity, overcome traditional time-series neural network to predict the shortcomings in the sequence cyclical lost. Experimental results show that, compared with conventional time series model and the traditional neural network model is simple, predicting ability and accuracy also improved.