Closed Atulsingh92 closed 1 month ago
I am not sure at this point if you can get explicit form for this quantity immediately from emukit, but it should be possible to estimate it. Assuming scale
is known, as you said, one can estimate mean and variance in any x at both fidelities, and then corresponding quantity is high - scale * low
. Tutorial notebooks on MF show to do predictions for individual fidelities.
Would that be good enough for you?
Answered with no further feedback for two years, closing. Please reach out again if necessary!
Hello, I wanted to know if there is a way to obtain this d_s(x) or the difference model for the equation shown here. The scaling parameter I am aware is available from one of the values within
lin_mf_model.gpy_model.param_array
command, however can't seem to find the difference values. (This is seen in code from Forrester's book in Matlab.) How can I get this d_s(x) values within emukit?