Description
Describe the feature or idea you want to propose
after discussion
this is in two parts. First is to allow the capability in the base class. This will require the introduction of a new tag, and a check of capabilities.
Then need to adapt the regression wrapper to allow exogenous variables.
Describe your proposed solution
Suppose your series to forecast is (1000,) and you want to use two exogenous variables stored in array of shape (2,1000). You want to use a window length 100. So if you have a horizon of 1, to forecast point 101, you use points 1 to 100 of the original series, and (2,100) of exogenous.
This makes a single case of shape (3,100) to predict one value of y.
A collection is then all of windows so created. You form a collection of shape X=(900,3,100), y=(900,).
Train/test split the same as if no exogenous, this is now just multivariate regression. Need to check the regressor is capable and may need tags.
Describe alternatives you've considered, if relevant
it may be useful to generalise it for variable length window, i.e. make it unequal length regression. When splitting the data, each case is all the points prior to forecasting point.
Additional context
No response