You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Is your feature request related to a current problem? Please describe.
I want to use two consecutive global forecasting model - I need fast inference times- back to back where the first predictions are treated as future covariates of the second model.
Describe proposed solution
A global ensemble model, that takes a list of global forecasting model and takes another global forecasting model as the regressor, but it should have save and load functionality and interfaces for interacting with each model parameters.
Describe potential alternatives
An alternative could be that global forecasting model that accepts a future covariates accept a model with the time series, and if supplied it uses that global model to first predict the required length of future covariates and consume them behind the scenes.
Additional context
We need a feature to enable us to use [stacking, boosting] concepts with global forecasting models as future covariates.
The text was updated successfully, but these errors were encountered:
Is your feature request related to a current problem? Please describe.
I want to use two consecutive global forecasting model - I need fast inference times- back to back where the first predictions are treated as future covariates of the second model.
Describe proposed solution
A global ensemble model, that takes a list of global forecasting model and takes another global forecasting model as the regressor, but it should have save and load functionality and interfaces for interacting with each model parameters.
Describe potential alternatives
An alternative could be that global forecasting model that accepts a future covariates accept a model with the time series, and if supplied it uses that global model to first predict the required length of future covariates and consume them behind the scenes.
Additional context
We need a feature to enable us to use [stacking, boosting] concepts with global forecasting models as future covariates.
The text was updated successfully, but these errors were encountered: