This message was deleted.
# neural-forecasts
Slackbot
10/02/2023, 2:11 PMThis message was deleted.
c
Cristian (Nixtla)
10/02/2023, 2:31 PMHi @Yalemzewd Nigussie. The horizon defines the number of points to forecast. If you are forecasting 8 values, you need to pass 8 data points in the
futr_dfunique_idCristian (Nixtla)
10/02/2023, 2:34 PMWe specify it in terms of the horizon (in points) because it does not depend on the sampling frequency. The idea is that for each value you are forecasting you need to have information of the future exogenous.
y
Yalemzewd Nigussie
10/02/2023, 2:46 PMso, H and len (futr_df) should not necessary be the same?
c
Cristian (Nixtla)
10/02/2023, 2:51 PMh should be the same as
len(futr_df)/df.unique_id.nunique()Cristian (Nixtla)
10/02/2023, 2:52 PMyou need h values for each time series
Cristian (Nixtla)
10/02/2023, 2:53 PMbut not any h values, they have to correspond to the exact datestamps you are forecasting. As we show in the tutorial
Cristian (Nixtla)
10/02/2023, 2:54 PMlets say you have the series with historic dates 1,2,3,4,5,6,7,8, and you are forecasting 9,10,11,12. You need to pass the values of the future variables for 9,10,11,12 dates.
y
Yalemzewd Nigussie
10/02/2023, 2:57 PMIn this example you provided, h=4, but I passed h=12. I got forecast results. There should be an error message to pop up right?
c
Cristian (Nixtla)
10/02/2023, 3:00 PMYes. We recently added a protection for this case, but it is currently only in the main branch of the repository. It is not yet in pip
y
Yalemzewd Nigussie
10/02/2023, 3:00 PMWhen, you plan that into production?
c
Cristian (Nixtla)
10/02/2023, 3:35 PMIt will be released this week, along other improvements
👍 1
Cristian (Nixtla)
10/02/2023, 9:09 PMThis is the error message you will receive
y
Yalemzewd Nigussie
10/03/2023, 12:02 AMThanks, for the update
4 Views