Open khoopes opened 4 years ago
You will find some answers in the paper (the link is broken in the SMT documentation).
extra_points is the number of additional nearby points that are sampled to make a first order Taylor approximation.
when should I enable them? The power of GEKPLS lies in using these extra points to improve predictions. So always adding a few extra points is good.
What are extra points in the context of GEKPLS? The documentation simply says "Number of extra points per training point" What is an extra point and when should I enable them?
Thanks,
Kevin