The Prediction Of Chinese Export Based On SARIMA Models And State Space Models

China’s foreign trade has a very important role in improving people’s living standards, developing the national economy and expanding the channels of employment. So it is very important to analysis foreign trade and to forecast Chinese import and export volume.This paper firstly introduces the significance and effect of foreign trade on China’s economic and social development, the status of China’s foreign trade and the proposed prediction models for the Chinese import and export volume, and then introduces the time series models and state space models. Chinese import and export volume, which were predicted by many scholars with the time series ARIMA model, have the characteristics of tendency and periodicity. Thus, this paper uses the seasonal ARIMA(SARIMA) model to improve the prediction accuracy. Due to the non-observability of the trend and seasonality of time series, this paper proposes to use the state space model of the control theory to tackle this. According to the characteristics of China’s export data, two state space models are established:the state space model1and the state space model2. Moreover, Kalman filtering algorithm is employed for the estimation of state parameters.The prediction results show that the SARIMA model and the state space models obtain good performance in the prediction of the Chinese export volume. The proposed models have good abilities such as strong reliability, validity and practicability, etc. And the state space model2outperforms the SARIMA model in forecasting precision.