Network Traffic Prediction Based On Support Vector Regression

With the expansion of internet-scale, the network management becomes more and moreimportant. Traffic of network is an important factor of network running status; it can recordand reflect the activity of users. So, the prediction of network traffic could provide aneffective basis for network bandwidth allocation, flow control, routing control, admissioncontrol, security management and so on.The Support Vector Machine (SVM) is a new and promising classification and regressiontechnique proposed by V. Vapnik based on statistical learning theory. As a standard kernellearning algorithm, it has successfully been used in image processing, text classification andbiological information processing, and also in time series prediction.In AdaBoost algorithm, for those samples who makes classifiers easily to make wrong pointsof classification are learned and vote for a second time. According to the result, the group ofstronger consistency will be choosed as the foundation of the integrated study. Therefore, therepeat study of the “high error area” is positive and obvious. Inspired by Adaboost, anensemble strategie is proposed in this paper that a further focused learning based on initialstudy:(1) Using LS-SVR in preliminary training: in a number of experiment and comparativeexperiments, the results show that the curve of flow is accurate, it is able to reflect the targetmovements, and training time is short, the limitations is that the distortion of details isobvious, thus, uses a single LS-SVR learner as a base is suitable, and there is space of levelup.(2) Based on the fitting error of initial training by LS-SVR, we use SVM classifier to findareas (high error area) which need to be focused learning.(3) A weighted based voting algorithm is applied in the ensemble learning, the results of thestudy will replace to the corresponding position.The experimental results show that, the study strategy of this paper is effective, combinedwith the advantages of single learning and ensemble learning, it makes a balance between thespeed and accuracy of training.