Prediction Of Passenger Railway Volume Based On ARIMAX

The influencing factors and theories of several kinds of prediction of railway passenger traffic volume are studied in this paper. On this basis, firstly, two single methods without considering the impact of factors are employed, including Box-Jenkins method and gray system method. Secondly, an aggregated model is proposed which combines gray system method with linear regression method. Subsequently, we select the significantly affected factors of railway passenger traffic volume by multiple regression analysis and stepwise regression, including GDP, national railway operating mileage and the total number of domestic tourism. Finally, we set these sequences as the sequence of input variables and passenger volume sequence as a response variable sequence to propose ARIMAX model. The differences between different prediction methods above are compared and their advantages and disadvantages are analyzed.Empirical analysis shows that prediction accuracy of the models considering impact of factors including ARIMAX model and gray linear regression combination model is higher than the single model without considering(either Box-Jenkins model or gray system model). The predicted effect of ARIMAX model is better than gray linear regression model and its result is less volatile and has a higher value of practical application.