An empirical study on GPRS traffic forecasting based on chaos and SVM theory

Forecasting of general packet radio service (GPRS) traffic directly influences the future development of 3rd-generation (3G) ervice of telecommunications enterprises. According to the complexity and non-linearity of GPRS traffic. In this paper, Firstly the schaotic feature of GPRS traffic is verified with chaos theory. It can be seen that GPRS traffic possesses chaotic features, providing a basis for performing short-term forecast of GPRS traffic with the help of chaos theory. The average mutual information (AMI) method is used to find the optimal time lag of the series. Cao's method is adopted to determine free parameters of support vector machines. Additionally, the proposed model was tested on the prediction of GPRS traffic of Chongqing province telecommunications operator in China to prove the model's validity.