How to predict furure values in a non linear regression?
1 view (last 30 days)
Torsten on 24 May 2021
Edited: Torsten on 24 May 2021
If your fitting curve depends on time t, evaluate it at the future time instant you have in mind.
But be very,very careful here:
Extrapolation of trend curves is extremely error-prone. Especially polynomial curves of degree greater 1 usually don't reflect the physical or economical background where the data stem from and are therefor only suited to interpolate between the time limits of your data.