search

SMTorg / smt

Surrogate Modeling Toolbox
https://smt.readthedocs.io/en/stable
BSD 3-Clause "New" or "Revised" License
648 stars 199 forks source link

RTMS explanation #213

Open ludoro opened 4 years ago

ludoro commented 4 years ago

Hello guys! First of all, really cool library.

I have been working on Surrogates.jl, a Julia package that mostly has the same goal. I am trying to implement from this paper here: "A fast-prediction surrogate model for large datasets" which I believe it's related to this library. However, I am not understanding how to build the matrix $F$, for each i-esim row I should evaluate the i-esim training point x at some splines, right? The number of columns is the degree of the splines? I am bit confused. Could you shed some light?

relf commented 3 years ago

Hi, thanks for the kind word. You're right the surrogate is related to this paper (link broken in the doc) and well... as far as I am concerned I am not that familiar with this surrogate so I would suggest to refer to the implementation... but maybe you already did?!

ludoro commented 3 years ago

Yeah, I tried but without much success. I get stuck at page 6 of that paper, when defining the F function. I am not in a hurry so I can maybe wait for someone who is familiar with it :)