The standard method
After we wish to predict future values for a time-series, we are sometimes fascinated about a number of future horizons, e.g. what’s going to occur in 1, 2, or 3 months. The standard method to foretell these totally different horizons consists in coaching a separate mannequin for every goal horizon.
A typical various consists in coaching a single mannequin on a brief horizon, and prolong it to multi-horizons by making use of it recursively (i.e. by taking the earlier predictions as inputs to provide the next ones). Nevertheless this method could be complicated to implement in manufacturing techniques, and it could result in error propagation: an error on a detailed horizon can have detrimental results for the next ones.
One other various consists in forecasting all of the horizons on the similar time with a multi-variate mannequin. Nevertheless, the sort of fashions that assist multi-variate outputs is restricted, and it requires additional effort in knowledge dealing with and mannequin upkeep.
Horizon as a function
A less complicated method consists in concatenating the info ready for every horizon, and including a brand new “horizon” function. This method has a number of benefits:
- It’s easy to grasp and implement, because it results in a single mannequin to coach and keep.
- It doubtlessly improves the predictions accuracy, for the reason that mannequin is skilled on a bigger dataset. It will possibly even be used as a “knowledge augmentation” approach: in case you are fascinated about only some horizons, you’ll be able to nonetheless add extra ones within the coaching part to enhance mannequin estimation.
- The mannequin can be utilized to foretell horizons on which it was not skilled, which is perhaps useful in case you have many horizons to foretell.
This method is the alter-ego of a international mannequin, however within the context of a number of horizons as an alternative of a number of…