The Mid Short Term Prediction Of Railway Passenger Traffic Volume In China

The short-term forecast of China's railway passenger traffic is the foundation of the railway company to develop annual plans,as well as the rational allocation of resources,expand the important basis for the work of passenger transport.In view of the development of air transport and passenger transport market competition,the railway passenger transport sector must be an early response to the changes in passenger traffic within the market to react in a timely manner.This paper predicts the railway passenger volume from the two levels of the annual forecast and monthly forecast,respectively.Both providing the decision support for the long-term development of railway transport and providing a reliable basis for the railway departments timely response to the changing market.In the forecast of the annual data,we proposes the M-CPSO-GM model based on the optimal model selection,and has a very good effect in the annual forecast of China's railway passenger traffic volume.In the forecast of the monthly data,we get the first step predictive value according to the S-ARIMA model firstly,then put it into the spring months and other months for processing.For the spring months,this paper puts forward an Spring Festival factor which is highly correlated with the one step prediction error as correction factor to correct the prediction values through fuzzy neural network.For non-spring months,we use the growth rate of last month as the correction factor,corrected by BP neural network.The results show that the prediction accuracy of the corrected monthly data has been significantly improved.Finally,the work of this paper is summarized,and the problems to be improved in this paper are briefly described,and the direction of the future of railway passenger traffic volume prediction in China is pointed out.