The Application Of Time Series Model In Internet Access Analysis In China

With the rapid development of science and technology, the Internet has the unprecedented development afterwards, the development of the Internet has played a great role in promoting the change of our way of life. The Internet has greatly promoted the development of online shopping, online shopping has stimulated the development of consumption, and has played a great role in promoting the economic growth, the Internet has become a new momentum of economic growth. At the same time, the development of the Internet has also promoted the innovation, and played an irreplaceable role in steady growth, structural adjustment and promote employment. The present Internet penetration of our country has reached nearly half, but in the next few years, it is an unknown number for the Internet penetration rate. Therefore, it is very important to predict the future of Internet penetration rate.Time series model is the most commonly used model for processing time series data. The time series model includes two kinds of models, one is based on the stationary time series, the other is based on the non-stationary time series. Although we often see non-stationary time series in practice, but the stationary time series model is the basis of non-stationary time series model. The development of building non-stationary time series model which based on stochastic analysis is based on stationary time series model. In the process of building non-stationary time series model, ARIMA model and Holt-Winters filter model is respectively based on stochastic analysis and deterministic analysis. The ARIMA model is mainly for the sequence which has relatively strong random nature; the Holt-Winters filter model is mainly for the sequence which has relatively weak random nature. Because we are unknown to the sequence before building the model, we usually use the two models to fit the non-stationary time series, and then compare and select the appropriate model.In this paper, we first introduce the principal character of time series and the steps of time series analysis, then introduce how to build the model of stationary time series, after that we analysis the modeling of non-stationary time series in the basis of the modeling of stationary time series. In the process of analysis, the modeling of non-stationary time series is divided into two kinds, one is based on deterministic analysis and the other is based on stochastic analysis, and the two kinds of modeling are compared to each other. Finally, we use the ARIMA model and Holt-Winters filtering model to fit the Internet penetration rate data in China from December 2000 to June 2015, in the basis of R software. Finally, after the model comparison, the Holt-Winters filter model is selected, and the Internet penetration rate in China in the next 5 years is predicted by the Holt-Winters filter model. The analysis shows that, China’s Internet penetration rate will reach 59.4% by June 2020,the increment speed is not very fast, so this paper puts forward the relevant policy recommendations to improve the growth of Internet penetration rate.