This message was deleted.
# general
s
This message was deleted.
k
I think you can just clip it with postprocessing?
m
@Thiago Vidigal, thanks for the question. Currently the only alternative is to do as @Kevin Kho proposes. However, if you want us to include it as a feature in
StatsForecast
you can open an issue here: https://github.com/Nixtla/statsforecast/issues/new/choose
Other alternatives are to use transformations on the data before training (like log transforms.)
t
Tkx for the aswers, I'm already doing like kevin said, it works.