Diagrams showing the development of values enable the user to compare track conditions of different periods. Consequently the planning and execution of reconditioning tasks is carried out more exactly. Therefore, these tasks work against aberrations more efficiently. (...)

The adjoining example displays the initial condition of an infrastructure in spring 2005 and a predicted condition for the end of 2007.
The used relation of predicted period to applied period of training data for the artificial neural network is about 1:3. That means, the performed prediction over 2,5 years needs a data stock of approximately 7,5 years. The standard deviation over 500 m of the super elevation was used. (...)