Forecasting: What an OR approach can do for you
12 May 2017
The newest OR Society Impact magazine contains an article by John Boylan on the topic of forecasting from an operations research perspective.
Professor John Boylan emphasises the importance of correct forecasting process in supply chain context focusing on three core topics:
First, he points out that the judgmental adjustments to statistical forecasts (which are very common in practice) do not necessarily increase the accuracy of forecasts. Furthermore, we already know which of the adjustments on average lead to increase or to decrease in the forecasting accuracy.
Second, John emphasises the importance of understanding the forecasting principles in practice. This means that the correct forecasting method should be selected for the data, meaningful parameters need to be chosen and the user should understand what to expect as a forecast from a method.
Third, he discusses the importance of assessing improvements in forecasting accuracy. It is not enough to say that the error has decreased with a new method – an appropriate comparison with benchmark models using appropriate error measures should be carried out.
Finally, John concludes that it is important to think in terms of financial benefits from accurate forecasts and to collaborate with O.R. and analytics specialists to see these benefits clearer.
The full article can be accessed for free.