Research on Construction of High-Efficiency Network Traffic Prediction Model in Edge Computing Environment

This study focuses on the network traffic prediction problem in edge computing environment, and designs a PSO-SVM integrated prediction framework, which aims to improve the prediction accuracy and response speed to meet the real-time and resource limitation requirements in edge computing environment. In the proposed framework, PSO algorithm is used to optimize the key parameters of SVM model. For example, the penalty coefficient <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{C}$</tex> and kernel function parameter <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\gamma$</tex>, as a bionic optimization algorithm, can quickly find the optimal solution in the parameter space through the cooperation and competition between particles, and then through the precise tuning of SVM parameters, the model not only improves the accuracy of prediction, but also enhances the generalization ability of the model. At the same time, SVM itself is a powerful supervised learning model, especially good at dealing with small sample data sets, and can effectively avoid overfitting. Combined with PSO and SVM, it can automatically adjust to the prediction of different network traffic patterns, so as to achieve fast and accurate network traffic prediction on edge devices. This innovative method provides strong support for the dynamic allocation and optimization of network resources, opens up a new path for the future development of edge computing and intelligent network technology, and is expected to promote the innovation of real-time network management and intelligent operation and maintenance technology.