Can one use both future and past covariates in sktime? #1809
Replies: 1 comment 1 reply
-
apologies for the late reply, I've missed this over the holidays. The Regarding the lagging of different variables ("only lag the variables not available in the future), you can use the The current prototype (not merged yet) for the lagged window transformer can be applied to sets of variables separately: |
Beta Was this translation helpful? Give feedback.
-
Would like to make use of past (observed variables) and future covariates (forecasts that are available before a forecast is made for that horizon). Examples:
sktime abstracts the creation of lagged variables, etc, so any way to make the distinction between the two types of covariates? In the example above,
temperature_forecast
should not be lagged, as it is available ahead of prediction time if only doing day ahead sales forecasts.observed_temperature
, however, should be lagged as the day needs to be completed to have the actual value. If both of these variables are used as exogenous inX
, can we make the distinction in sktime?Beta Was this translation helpful? Give feedback.
All reactions