The topic of forecasting adhesion lends itself well to novel data analytics techniques as the correlating factors that determine the railhead condition are very complex and it is often difficult to understand them using traditional techniques.