Predicting rail geometry deterioration by regression models

In this paper we analyse railway track geometry deterioration data on characteristics like scant, horizontal, vertical alignment, as provided by the Eurailscout measurement train. We apply statistical regression techniques to assess the influence of environmental characteristics, like subsoil, tonnage and underlying objects. We also assess the direct quality improvement because of maintenance. The results are used to make a forecasting model for deterioration and to assess the amount of maintenance necessary to maintain track quality.