Keynote talk given at International Symposium on Forecasting 2020
It is common to forecast at different levels of aggregation. For example, a retail company will want national forecasts, state forecasts, and store-level forecasts. And they will want them for all products, for groups of products, and for individual products.
Ten years ago, anyone doing such forecasts needed to select between bottom-up, top-down or middle-out methods. Then optimal forecast reconciliation was introduced, and a new and better approach was available. In this approach, forecasts at all levels of aggregation are produced, and the results are “reconciled” so they are consistent with each other.
In the last ten years, the literature on forecast reconciliation has exploded with developments in theory, methods, applications and software. I will provide an up-to-date overview of this work and show how reconciliation methods can lead to better forecasts and better forecasting practices.
I will also show how these ideas can be easily implemented using the fable package in R.