The editorial board of the International Journal of Forecasting awarded the publication entitled “Improving forecasting by estimating time series structural components across multiple frequencies” with the award for publications in 2014 and 2015.

The authors propose a new forecasting approach which makes use of information available at different levels of temporal aggregation. The method not only provides forecast performance improvements but also mitigates model selection problems by making use of forecast combinations. A further advantage of the approach is that it leads to aligned short, medium and long-term forecasts, helping to make better decisions at all planning levels. The method is available as an R package (MAPA) and is described in further details on Nikos' blog.

During the award ceremony at the ISF2017 conference, editor-in-chief Professor Rob Hyndman said: “This paper tackles a very important and difficult forecasting problem, combining forecasts when data are available at different frequencies. It succeeds by finding a forecast procedure robust to model selection and parameterization while being simultaneously simple to implement. As a consequence of this simplicity, the procedure can be important not only for academics but also for practitioners.”

The Centre for Marketing Analytics and Forecasting, of which all three are members, congratulated the authors on their great achievement.