Empirical Study On The Choose Of Forecast Model For Railway Freight Turnover

According to the statistics of China Railway Statistics Compilation (2006), the demand for transport over exceeds its supply. Among various trasportation mathods, railway is the biggest problem. It has become the bollteneck of GDP development. The reform and development of Chinese railway is put on the schedule. However, we must be clear that the development should not be blindfold. The development should based on the study of railway economy theories. A precise forecasting of the railway freight turnover is the key to analysis the freight demand-supply relation, make sure the development direction, and establish future develop statistics of railway trasport. However, regarding to the complexity of railway freight transport: mixed transport of freight and passengers, the railway net can not satisfy the real demand, and freight turnover data can not be collected, the forecastings study of railway turnover should be continued and developed. Now, we are facing great changes: the separate of freight and passenger line, the develop of city-city passenger transport system, and the new speed increase. Setting transport policy, defining develop direction, and sloveing freight forcasting problem are put forward once again.Regarding to the situation and problems I mentioned above, this paper chooses to study on the forecast of railway freight turnover; intend to put forward a feasible way to maintain relavent issues; try to slove the problems in freight forecast by studying forecast methods, investigating and analyzing historical turnover data, situation and experience; find the rule and trend of freight transport demand; establishing the foundation of freigt transport plan and decision making. The main job of this paper is as following:1. Collection & Summary of forecasting methods. There are plenty of forecasting methods. In this paper, I separate them into three classes: relative factor analysis, time series analysis and other forecasting methods, and compare them based on their develop process, model choice, coefficient estimation, as well as their advantages and short comes.2. The current situation and develop planning of railway freight transport. As the study object of this paper, railway freight turnover forecast has characteristics defferent from other objects. Only through a over all look at these characters, can we have a porper analysis on freight turnover data, and establish an effective model.3. A study on actual-fitted model out of historical freight turnover data. Establishing freight turnover actual-fitted model, and analyze its fitting efficiency. Demonstration forecast of freight turnover forecasting, based on the fitted models, which have pased the tests.