Abstract: In order to improve the network traffic prediction accuracy, this paper proposes a network traffic prediction method based on least sctuare support vector machine(LSSVM) optimized by genetic algorithm which uses the relation between phase space reconstruction and parameters of prediction model. Firstly, phase space reconstruction and the parameters of LSSVM were used as an individual of genetic algorithm while the model prediction accuracy was used as the fitness function, and then global optimal parameters of the model were obtained by genetic algorithm, lastly, the simulation tests were carried out based on network traffic data. The results show that, compared with the traditional forecasting methods, the proposed model improves the prediction accuracy of network traffic and provide a new research thought for network traffic prediction.

Key words: Network traffic, Phase space reconstruction, LSSVM, GA