Analysis and Forecast of Railway Freight Ton-Kilometers Based on Partial Least Square Regression

Railway Freight Ton-Kilometers (RFTK) mainly results from the interaction between the railway freight transportation demands (RFTD) and the effective supplies of the railway transportation capacity. Based on the qualitative analysis of the influence factors of RFTK, some variables are selected as the influence factors reflecting the supplies of railway transportation capacity. They are the length of railways in operation, the length of electrified railway, the double-tracking length weight, the automatic blocking length weight and the running speed of freight, trains; other variables are selected as the influence factors reflecting the situation of the national economy. They are GDP, the industry production value, the coal output, the total value of imports and exports. The model is built by the means of Partial Least Square Regression (PLS). Modeling is realized by the software SAS 9.0. The results indicate that: (1) the model built by PLS is effective and feasible in this problem (2) In recent years, although the extension scale in the railroad network of our country is not small, it can not match the rapid economic growth; In addition, the running speed of freight trains has been slow for a long time. These are all the restrictive factors of the increase of RFTK. The enhancement of the railway transportation capacity in our country has to depend mainly on the increase of the level of railway technical equipment modernization.