What is ensemble forecasting in the Automatic Best Method forecast method?
Ensemble is a process that blends various forecasting models into one forecast. If the Automatic Best Method forecast is always generating a forecast using ensemble, it is not always the same underlying models. For example, ensemble run one could use a blend of Holt Arima and Walking average. Ensemble run two could use a blend of Theta and Point Estimate Weighted Average.
Currently, all the underlying models in ensemble must have the same weighting. For example, if two models are selected, each one is weighted 50 percent. If four models are selected, each one is weighted 25 percent. In a future update, we will display the underlying models and weight in an ensemble forecast.
For more information, see ensemble forecasting.