An Application of Dynamic Regression Model and Residual Auto-Regressive Model in Time Series

This paper, by using the dynamic regression model (ARIMAX) models and predicts the tourist date from 1979 to 2004 in Zhejiang Province, and makes stationary test and white noise test of the residual date generated by the above analysis . The innovation point of this paper is that it is suitable to establish dynamic regression by cointegration test and proves the data of the residual data validation is stationary. Further testing and analysis of residual data, finds that the residual data can establish auto-regression model. This method has made full use of data information. Thus the paper presents that the prediction effect of the combination of the dynamic regression model and the residual autoregressive model is superior to that of the prediction model of the ARMA model. This combination model has better adaptability, greatly improves the predicted effect of the model.