LS-SVM based Prediction of the Size of Internet User in China

摘要

It is very useful to predict the growth of internet user in China. A least square support vector machine regression model was introduced. Firstly, phase space reconstruction technology was used to deal with time series data, then with LS-SVM, a method to predict the number of Internet user was constructed. The numerical experiments show that the accuracy of this prediction model has great advantage compared with the Grey theory and linear regression-based prediction model. It is a good alternative to predict the growth of Internet in China.