Network traffic prediction based on least squares support vector machine with simple estimation of Gaussian kernel width

In order to improve the accuracy of network traffic prediction and overcome the disadvantages of slow convergence speed and easy to fall into local minimum value in the process of least squares support vector machine (LSSVM) network traffic prediction, a network traffic security prediction model based on LSSVM which simply estimates the width of Gaussian kernel is proposed. The model assigns different Gauss kernel widths for each sampling point according to the local density of the sampling point. The simulation results show that, compared with LSSVM and PSO-LSSVM, the model proposed in this paper improves the accuracy of network traffic security prediction, reduces the training time of sample data, and provides strong decision support for network traffic planning and network security management.