Currently, the gap filling requires a leading and trailing period. This is needed for the assignment of weight used in the debiassing.
Proposal by @amberJ99 is to also implement a method without weights (so leading/trailing becomes one learning period).
This can be done by adding an extra argument or something like that.
Currently, the gap filling requires a leading and trailing period. This is needed for the assignment of weight used in the debiassing.
Proposal by @amberJ99 is to also implement a method without weights (so leading/trailing becomes one learning period). This can be done by adding an extra argument or something like that.