You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am learning a model on experimental data for which the trend is well known.
The data has some big regions of missing data.
The mean function is important so the model doesn't go to zero when we are inferring inside the missing data regions.
The issue I have is that the current mean functions do not allow for composition in the way kernels do. In particular one cannot select the dimension on which the mean function is active. That is, afaik, this mean function cannot be defined, although the component functions are available in gpflow:
Y= (a*X[:,0] +b) +X[:,1]**2
which would be something like the following if dimensions could be selected:
What's the gpflow way of achieving this?
I couldn't find a concise definition of the interface for mean functions, specially if I want to write ones that accept active_dims.
Any help is appreciated.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
Hi all,
Thanks for the software.
I am learning a model on experimental data for which the trend is well known.
The data has some big regions of missing data.
The mean function is important so the model doesn't go to zero when we are inferring inside the missing data regions.
The issue I have is that the current mean functions do not allow for composition in the way kernels do. In particular one cannot select the dimension on which the mean function is active. That is, afaik, this mean function cannot be defined, although the component functions are available in gpflow:
which would be something like the following if dimensions could be selected:
What's the gpflow way of achieving this?
I couldn't find a concise definition of the interface for mean functions, specially if I want to write ones that accept active_dims.
Any help is appreciated.
Beta Was this translation helpful? Give feedback.
All reactions